Geology shapes biogeography: Quaternary river-capture explains New Zealand's biologically ‘composite’ Taieri River

Quaternary Science Reviews 120 (2015) 47e56

Contents lists available at ScienceDirect

Quaternary Science Reviews journal homepage: www.elsevier.com/locate/quascirev

Geology shapes biogeography: Quaternary river-capture explains New Zealand's biologically ‘composite’ Taieri River Jonathan M. Waters a, *, Graham P. Wallis a, Christopher P. Burridge b, Dave Craw c a

Department of Zoology, University of Otago, PO Box 56, Dunedin 9054, New Zealand School of Biological Sciences, University of Tasmania, Hobart 7001, Tasmania, Australia c Department of Geology, University of Otago, PO Box 56, Dunedin 9054, New Zealand b

a r t i c l e i n f o

a b s t r a c t

Article history: Received 16 February 2015 Received in revised form 17 April 2015 Accepted 28 April 2015 Available online

Geological processes are hypothesised to strongly affect species distributions. In particular, a combination of geological and biological data has suggested that tectonic processes can drive vicariant isolation and speciation in freshwater-limited taxa. Here we synthesise geological and biological evidence to demonstrate a composite geological and biological history for New Zealand's 290-km long Taieri River. Specifically, we assess evidence from structural geology and petrology, combined with phylogenetic and biogeographic analysis of galaxiid fishes, to show that the modern Taieri River was formed via capture of the ancestral Kye Burn during the mid-late Quaternary. Molecular dating analyses support a lateQuaternary timeframe for the geologically-mediated divergence between formerly-connected sister taxa Galaxias depressiceps and G. ‘teviot’. Fish biogeography lends further support to the geological hypothesis, as there is a substantial biogeographic disjunction between the lower- (ancestral) and upper (captured) portions of the Taieri River. Geological and biological data are assessed independently yet yield consilient patterns and timeframes for the evolutionary events inferred. Broadly, this study highlights the interplay between physical and biological processes in a geologically dynamic setting. © 2015 Elsevier Ltd. All rights reserved.

Keywords: Biogeography Dating Evolution River capture Tectonics Vicariance

1. Introduction One of the primary goals of biogeography is to understand the biological and physical processes underpinning the evolution and distribution of species around the globe (Briggs, 1995; Lomolino et al., 2005; Gillespie et al., 2012). Freshwater systems represent particularly fertile systems for studying interactions between geological processes (Bishop, 1995) and biological evolution ~rescu, 1990). Specifically, freshwater-limited (Mayden, 1988; B~ ana populations tend to have relatively restricted geographic ranges (McDowall, 1990; Berra, 2001; Leathwick et al., 2008), are particularly prone to vicariant isolation (Mayden, 1988; Matthews, 1998; Near and Keck, 2005; Burridge et al., 2006; Kozak et al., 2006) and often exhibit contrasting levels of genetic divergence within and among catchments (Avise et al., 1987; Ward et al., 1994; Avise, 2000; DeWoody and Avise, 2000). Indeed, genetic analyses from several parts of the globe have recently revealed tight links

* Corresponding author. Allan Wilson Centre, Department of Zoology, University of Otago, PO Box 56, Dunedin 9054, New Zealand. E-mail address: [email protected] (J.M. Waters). http://dx.doi.org/10.1016/j.quascirev.2015.04.023 0277-3791/© 2015 Elsevier Ltd. All rights reserved.

between landforms and associated freshwater biotas [e.g. (White et al., 2009; Goodier et al., 2011; Schwarzer et al., 2011; Gottscho, 2015)]. New Zealand is a geographically isolated and geologically dynamic region that presents a particularly informative setting for assessing relationships between geological history and biological evolution (Fleming, 1979; Goldberg et al., 2008; Wallis and Trewick, 2009; Heenan and McGlone, 2012). The country's recent history of tectonic uplift has had particularly profound evolutionary effects on freshwater-limited taxa (Wallis and Trewick, 2009). In particular, phylogeographic studies of New Zealand's freshwater species have revealed numerous cases of population fragmentation, with strong regional cladogenesis within and among taxa (Waters et al., 2001; Burridge et al., 2006; Apte et al., 2007; Craw et al., 2007a). The evolution and biogeography of New Zealand's distinctive galaxiid fish fauna has interested scientists for decades (McDowall, 1970, 1990). In particular, the species-rich Galaxias vulgaris complex has an intriguing widespread biogeographic distribution in South Island, spanning no fewer than 40 distinct catchments. While the group was once considered to represent a paraphyletic assemblage formed via multiple convergent losses of marine dispersal ability (Allibone and Wallis, 1993; Waters et al., 2001a), it now seems

48

J.M. Waters et al. / Quaternary Science Reviews 120 (2015) 47e56

likely that this radiation has a monophyletic freshwater origin (Waters et al., 2010), and that its biogeographic history has been shaped by river drainage evolution (Waters et al., 2001; Burridge et al., 2008b). Phylogenetic and statistical phylogeographic methods present powerful tools for reconstructing histories of recent-evolved species assemblages (Avise, 2000; Lemey et al., 2009). Such approaches can potentially be used to unravel biogeographic histories [e.g. (Funk et al., 1995; Shaw, 1996; Hewitt, 2000; Juan et al., 2000; Picard et al., 2008; Wallis and Trewick, 2009)], including inferences of ancestral taxon ranges, and the timings of dispersal and vicariant events. In the case of New Zealand's species-rich G. vulgaris complex, links between earth-history and freshwater biogeography are compelling (Waters et al., 2001; Burridge et al., 2008b). While detailed phylogeographic reconstructions of ancestral ranges for this assemblage have yet to be undertaken, extant species distributions alone suggest that the complex likely has an origin in southern South Island (Fig. 1), as the region boasts relatively high species-diversity (Allibone et al., 1996; McDowall and Wallis, 1996; McDowall, 1997; Wallis et al., 2001; Waters et al., 2001b). McDowall and Wallis (McDowall and Wallis, 1996), for instance, suggested that the Canterbury lineage (to the north) “may be derived from one of the southern forms … by dispersal north from the Taieri or Clutha”. This ‘out of the south’ scenario is also supported by regional contrasts in levels of allozyme (Wallis et al., 2001) and mtDNA (Waters et al., 2001b) variation, with consistently higher genetic diversity found in southeastern South Island (Otago) populations. Robust inferences regarding the role of geological processes in shaping biological phenomena rely on independent (non-circular) assessment of biological and geological evidence (Bishop, 1995). Previous structural geological analyses of central Otago have suggested that the region has a complex tectonic history that has driven substantial changes in drainage geometry from Miocene to present (Youngson et al., 1998; Craw et al., 2012; Upton et al., 2014). Similarly, the comparatively high species diversity of Taieri River Galaxias suggests a complex biogeographic history. We hypothesise that the 290 km-long Taieri River (Fig. 1a) has a composite history, reflecting mid-Quaternary uplift and associated river capture events. Here we synthesise existing geological and biogeographic data, and present new analyses, to elucidate the history of the Taieri River and test the hypothesis that drainage reversal has shaped fish biogeography and cladogenesis. 2. Geological setting 2.1. Initial development The Otago region of southern South Island, New Zealand (Fig. 1a,b) has a dynamic geological history. The basement rocks are Mesozoic schist and greywacke, and the area of interest for this study occurs entirely on the schist (Fig. 1a,b). Regional uplift of the basement rocks from beneath the sea began in the late Cenozoic (~23 Ma) with the inception of a new tectonic plate boundary, the Alpine Fault (Fig. 1a inset; (Cooper et al., 1987; Sutherland, 1995; Landis et al., 2008)). Uplift of mountains and hills was initially accompanied by Miocene volcanism at the east coast (Figs. 1b and 2; (Coombs et al., 1986; Hoernle et al., 2006). Localised subsidence resulted in development of a large lake complex, Lake Manuherikia, during the middle Miocene (Fig. 2; (Douglas, 1986; Marsaglia et al., 2011; Upton et al., 2014)). In the Pliocene, debris from rising mountain ranges filled the lake to create a broad braided river plain that drained to the ancestral Clutha River (Fig. 2; (Douglas, 1986; Youngson et al., 1998; Craw et al., 2012; Upton et al., 2014)). This debris was dominated by greywacke

from ranges to the northeast (Figs. 1b and 2). Paleocurrent indicators within these remnants confirm the south to southwestward drainage of that time (Fig. 2; (Youngson et al., 1998)). 2.2. Quaternary drainage reorientation due to folded schist mountain uplift The schist basement of the Otago region, including the Taieri River catchment, has a strongly developed planar schistosity that is generally nearly horizontal. This pervasive structural feature constitutes planes of weakness in the rock that permit the rock to fold during compressional deformation. Ongoing deformation associated with the Alpine Fault during the Quaternary has resulted in development of broad open folds of this schistosity to form antiform (upfolded) mountain ranges with intervening synform (downfold) valleys, with a wavelength of ~20 km (Figs. 1a,b; 2; 3a; (Landis et al., 2008)). Faults have developed in tighter portions of these folds (Fig. 3a; (Jackson et al., 1996; Bennett et al., 2005, 2006)). Initiation of the rise of these mountain ranges has been dated to the middle Quaternary (~1 Ma) via cosmogenic analysis, allowing estimates of uplift rates (Bennett et al., 2005, 2006). These structures are still active, and several active faults occur along their margins (Turnbull, 2000; Forsyth, 2001). 2.3. Evolution of the Taieri River catchment The most important features constraining the geometry of the Taieri River are the three antiform fold ranges that define what is now the upper Taieri catchment: South Rough Ridge, Lammermoor Range, and Rock & Pillar Range (Figs. 1a; 2; 3a,b). The Rock & Pillar Range has developed at the edge of the pre-existing Miocene uplift zone, and the other two ranges have developed from beneath the Pliocene river plain. An additional antiformal range, North Rough Ridge, has grown and impinged on to South Rough Ridge, and an adjacent range is currently emerging through the sedimentary cover to the northeast of South Rough Ridge (Bennett et al., 2005, 2006). The Lammermoor Range, which is oriented at a high angle to the other ranges, has been controlled by a fault, the Teviot Fault, on its southwestern margin (Figs. 1b and 3a). This fault, which locally controls the Clutha River course, is an old basement structure that became reactivated in the late Cenozoic (Turnbull, 2000; Craw et al., 2012; Upton et al., 2014). Growth of the antiformal fold mountains in the middle Quaternary was responsible for the formation of the present Taieri River drainage geometry. In particular, the rising South Rough Ridge, Lammermoor Range, and Rock & Pillar Range blocked the southward flow of a major Clutha River tributary, the Kye Burn (Figs. 1a; 2; 3a). As a result, this river was diverted to the southeast, into the ancestral Taieri River catchment, where it has since cut a deep gorge in that paleodivide, through a layer of Miocene lava flows and underlying schist basement (Fig. 3a). At the same time, the former middle reaches of the ancestral Kye Burn (now the upper Taieri) had their direction reversed, now flowing northeast along the edges of the growing Rock & Pillar Range before falling through the new gorge to the south (Fig. 3a). More locally, a previously southwest flowing stream at the southern end of South Rough Ridge became captured by northeast flowing Deep Creek (Taieri; Fig. 3a,b). 2.4. Taieri River catchment evolution in a wider context A drainage reversal event similar to that of the Taieri River described in the previous section, occurred in the middle reaches of the ancestral Ida Burn (west of the Kye Burn) (Fig. 1a). This reversal occurred as a result of growing antiformal fold ranges to the west of

Fig. 1. Location, geology, and topography of the Taieri River catchment and adjacent catchments in Otago, southern New Zealand. (a) Present topography. Principal drainages relevant to this study: CR ¼ Clutha River; TR ¼ Tokomairiro River; SR ¼ Shag River; KB ¼ Kye Burn; IB ¼ Ida Burn; PB ¼ Pool Burn; MR ¼ Manuherikia River; WR ¼ Waitaki River. Fold ranges relevant to this study are: R&P ¼ Rock & Pillar Range; LR ¼ Lammermoor Range; SRR, NRR ¼ South and North Rough Ridge respectively. (b) Geological map of the same area as in (a), after (Turnbull, 2000; Forsyth, 2001), showing the principal basement rock types and the distribution of Cenozoic cover rocks. Principal Quaternary folds of schist foliation are indicated.

50

J.M. Waters et al. / Quaternary Science Reviews 120 (2015) 47e56

Fig. 2. Paleotopography of the same area as in (a), based on geological inferences (see text; (Youngson et al., 1998; Craw et al., 2012)).

the Taieri catchment (Figs. 1a,b, 2, 3a). This reversal resulted in formation of what is now the northeast-flowing Pool Burn, which joins the modern Ida Burn before cutting through a deep gorge through an antiform range to the west (Figs. 1a; 2 (Jackson et al., 1996)). This drainage reorientation is essentially a mirror-image version of the diversion of the ancestral Kye Burn into the Taieri River catchment (Figs. 1a; 2). The Manuherikia River, which now receives the diverted Pool Burn and Ida Burn, is a remnant of the original Pliocene drainage system from the greywacke ranges (Figs. 1a; 2; (Craw et al., 2012)). Progressive growth of the South and North Rough Ridge antiforms has shaped a complex ridge that now forms a drainage divide between the enlarged Taieri River catchment and the Clutha River catchment to the west (Fig. 1a, b). This drainage-divide initially developed in the middle Quaternary (Bennett et al., 2005, 2006; Craw et al., 2007b) but was later breached by a small tributary of what are now the headwaters of the Ida Burn by progressive uplift of an adjacent antiform in the Taieri River catchment (Figs. 1a,b, 2, 3a; (Bennett et al., 2005, 2006; Craw et al., 2007b)). This breach resulted in a sharp westward change in stream orientation and associated cutting of a gorge through schist basement in the late Quaternary (Figs. 2 and 3a; (Craw et al., 2007b)). 3. Greywacke clasts in Pliocene to Holocene sediments

clasts contrast with locally-derived schist debris (Fig. 2). Likewise, paleocurrent directions deduced from outcrops of these Pliocene gravels provide support for this southerly drainage direction (Fig. 2). Much of the Pliocene deposits have been eroded during subsequent uplift, but some relict greywacke still persists in younger sediments. This record is extensive in the Manuherikia River catchment, where greywacke clasts are preserved in gravel deposits of numerous ages from Pliocene, through Pleistocene, to the Holocene (Fig. 2; (Craw et al., 2012)). However, erosion during uplift of the antiform fold ranges has removed greywacke clasts from what were the middle and lower reaches of the ancestral Kye Burn and Ida Burn courses (Fig. 2). Hence, evidence for past southeastward river flows towards the Clutha River is preserved mainly in the upper reaches of the ancestral Ida Burn and Kye Burn catchments in which only minor uplift has occurred (Fig. 2). Erosion of Pliocene greywacke clasts was facilitated by chemical decomposition of most such clasts by groundwater (Chamberlain et al., 1999). Water-rock interaction resulted in alteration of feldspars and micas in the greywacke, ultimately leading to rocks dominated by a range of clay minerals, and relict quartz (Chamberlain et al., 1999). This clay alteration caused the rocks to become soft and friable, resulting in disaggregation to clay and quartz sand during erosion, and subsequent removal of this material by rivers. Only quartz-rich rocks survived this alteration process (Youngson et al., 1998; Chamberlain et al., 1999). Similar processes persist, to a lesser extent, in Quaternary sediments (Chamberlain et al., 1999).

3.1. Differential preservation of greywacke-bearing clasts 3.2. Greywacke clasts in the Taieri catchment The extensive greywacke debris that was shed southwards from the northern greywacke ranges in the Pliocene provides evidence for paleodrainage directions in the schist belt, as the distinctive greywacke

The new drainage divide at the head of the modern Taieri River, on the Lammermoor Range, is a region of low rolling relief in which

J.M. Waters et al. / Quaternary Science Reviews 120 (2015) 47e56

51

Fig. 3. Topography of present upper Taieri catchment. A. Sketch block diagram (not to scale; vertically exaggerated) of geological and topographic features of the upper Taieri. The catchment topography is dominated by Quaternary folding (antiforms and synforms) of the schist basement, with some associated faulting. Rise of the antiforms has interrupted ancestral river flow directions (dashed arrows). B. Map of present drainage divide (as indicated in A) showing stream geometry, galaxiid fish distribution, and inferred paleodrainage directions. C. Photograph of a conglomerate pebble sampled from Teviot River catchment (as indicated in B) but originally derived from greywacke mountains at the head of the ancestral Kye Burn.

52

J.M. Waters et al. / Quaternary Science Reviews 120 (2015) 47e56

small pockets of Quaternary sediments have been preserved during uplift and drainage reorientation (Fig. 3b). These sediments contain alluvial gold, derived from the underlying basement schist and from remnants of Cenozoic river sediments, and have been mined historically, leaving excavational exposures. The sediments are dominated by schist debris derived from immediately adjacent areas, with a variable component of quartz conglomerate from the Cenozoic sediments. One of these sediment pockets, immediately west of the drainage divide at Lake Onslow also contains rare pebbles that have been derived from the greywacke ranges to the northeast, on the other side of the present drainage divide (Figs. 2 and 3b). These pebbles include quartz-rich (siliceous) argillite and one pebble of conglomerate (Fig. 3c), remains that have apparently survived chemical decomposition because they are relatively quartz-rich and highly lithified compared to most greywacke material. Despite the rarity of these pebbles, they provide direct evidence for ancestral flow of the Kye Burn across what is now a schist drainage divide (Fig. 3a,b). The modern Taieri River contains abundant greywacke-derived gravel and sand extending from its confluence with the Kye Burn to the sea. Typical modern gravels in the lower reaches of the Taieri River contain at least 20e30% greywacke pebbles. By contrast, older gravel deposits are rare in the Taieri catchment, apart from alluvial

fans that extend from the rising fold ranges, and these fans are entirely derived from local schist basement. Small remnants of Quaternary gravel of unknown age occur in abandoned river channels >50 m above the present Taieri River, ~50 km downstream of the present Kye Burn confluence (Fig. 1a; 2). These gravel deposits consist of schist clasts that have been extensively chemically decomposed, and no greywacke clasts have been found in them. This observation provides further evidence, albeit negative, for Quaternary bypassing of the ancestral Taieri catchment by drainage from the rising greywacke mountains to the north (Fig. 2). 4. Biological evidence for a composite Taieri catchment 4.1. Taieri fish biogeography Members of the freshwater-limited G. vulgaris complex have distinctive biogeographies (Leathwick et al., 2008), with strong phylogeographic structuring (regional clades) within and among taxa (Fig. 4). In addition, a combination of distributional (this study) and phylogenetic (previous studies) evidence suggests that the G. vulgaris complex in particular has a long evolutionary association with the Taieri River. First, the earliest split of the complex leads to Galaxias eldoni and Galaxias pullus ((Waters et al., 2010); Fig. 5),

Fig. 4. Phylogeographic relationships within the ‘flathead’ clade of the Galaxias vulgaris complex, showing the sister relationship between G. depressiceps and G. ‘teviot’. Phylogenetic relationships, molecular date estimates and nodal support values are from the chronogram of (Burridge et al., 2012), and major phylogeographic units within taxa are demarcated by dashed lines. These data represent a synthesis of several published analyses (Waters and Wallis,2001a; Waters and Wallis, 2001b; Waters et al., 2001; Burridge et al., 2006; Burridge et al., 2007; Burridge et al., 2008a; Burridge et al., 2008b).

J.M. Waters et al. / Quaternary Science Reviews 120 (2015) 47e56

53

Fig. 5. Distribution of Galaxias vulgaris-complex species within the Taieri River (catchment boundary indicated by dashed line). Species identities are indicated by colour on the phylogenetic tree (inset; after (Waters et al., 2010)). The distribution of G. ‘teviot’ is also shown (this lineage is absent from the rest of the Clutha system). Additional distributional data (not shown): G. pullus is present in lower Clutha and Tokomairiro tributaries (south of the Waipori); G. eldoni is also present in the Tokomairiro; G. depressiceps occurs in some adjacent coastal drainages; G. anomalus occurs widely in the Manuherikia (Clutha) system; G. ‘spD’ is widespread throughout the Clutha (Fig. 4).

both endemic to Otago, and is centred on the lower Taieri River (Fig. 4). Similarly, the most basal split within the ‘flathead’ clade leads to Galaxias depressiceps (Figs. 3 and 4; (Waters et al., 2010)), and is centred on the upper Taieri (Figs. 4 and 5). Third, in terms of Galaxias species-richness, the Taieri is unusually diverse, with four species (G. eldoni, G. pullus, G. depressiceps, G. anomalus) widely represented in the system ((Allibone et al., 1996; McDowall and Wallis, 1996; McDowall, 1997; Waters and Wallis, 2001b); Fig. 5). Additionally, it has two very narrowly distributed lineages (G. ‘teviot’, G. ‘spD’; Fig. 5), the latter (and perhaps the former) from localised anthropogenic disturbance (Esa et al., 2000; Waters and Wallis, 2001a). Apart from the larger Clutha system (Waters et al., 2001; Carrea et al., 2013), the Taieri is the only New Zealand river to house more than two members of this species complex (and few have more than one). We propose that this high diversity reflects geologically-mediated secondary contact of divergent lineages that originally evolved in separate catchments, but became subsequently connected via river capture (Figs. 2 and 6). New Zealand has an extensive database of freshwater fish locality records, with species identifications based on a combination of genetic and morphological criteria developed over the last two decades (Allibone et al., 1996; McDowall and Wallis, 1996; McDowall, 1997; Waters and Wallis, 2001b). An analysis of species-location records from the Taieri reveals striking biogeographic structure in the distribution of species, suggestive of a composite origin. Specifically, a search of this database [https:// nzffdms.niwa.co.nz/search] performed in January 2015, focussing on the Taieri River (catchment 743) retrieved 45 locality records for G. pullus (altitude range 400e1050 m a.s.l.), 94 records for G. eldoni

(range 139e1015 m), 141 for G. anomalus (190e785 m) and 172 for G. depressiceps (170e1280 m). These locality data indicate that G. eldoni and G. pullus together dominate the lower reaches of the Taieri, whereas G. depressiceps and G. anomalus together dominate its upper reaches (Fig. 5). It should be noted that, while the distributions of these taxa have apparently been locally fragmented by the introduction of invasive predatory and competitive salmonids that dominate most high-order streams in New Zealand (McDowall, 1990; Allibone et al., 1996), this issue is unlikely to have affected broad biogeographic partitioning of taxa (Figs. 4 and 5). It

Fig. 6. Schematic summary of mid-late Quaternary river drainage evolution and associated freshwater biogeographic changes in the Taieri River. Mid-late Quaternary capture of the Kye Burn by the Taieri River explains the composite biogeography of the modern Taieri system. Distinct faunal assemblages are illustrated with coloured regions (biogeographic details from Fig. 4 and 5), and geological details are illustrated in Figs. 1e3.

54

J.M. Waters et al. / Quaternary Science Reviews 120 (2015) 47e56

is also clear that all four galaxiid species exhibit broad altitudinal ranges (extending from lowland to subalpine/alpine) implying that altitude alone is not a primary factor explaining their disjunct distributions. Similarly, three of the four species have been documented both close to the coast (G. depressiceps occurs in additional coastal drainages), and substantial distances inland, so inland penetration alone seems unlikely to explain the disjunct distribution pattern. It is also unlikely that ecological differences among taxa have shaped this broad biogeographic structure. For instance, while there are clear ecological differences between G. depressiceps (high-gradient streams; (McDowall and Wallis, 1996)) and G. anomalus (low-gradient streams; (McDowall and Wallis, 1996; Leprieur et al., 2006)), these taxa retain broadly overlapping ranges in the upper Taieri (Fig. 5). Instead, we suggest that drainage history has played a key role in shaping biogeographic disjunction in this region. We propose that the distinct upper- (depressiceps, anomalus) versus lower-Taieri (pullus, eldoni) faunas reflect the river's composite geological origin (Fig. 6). Specifically, the divergent eldonipullus clade d originally endemic to the ancestral (lower) Taieri (Fig. 2) d likely came into contact with the upper Taieri assemblage (depressiceps, anomalus) only once the ancestral Kye Burn was captured by the lower Taieri. Despite many hundreds of thousands of years of river connectivity since the capture event, the original geological and biogeographic pattern has been preserved, a phenomenon similar to that previously noted in the Nevis-Clutha River system (Waters et al., 2001). 4.2. Teviot lineage and molecular dating of river capture A recent galaxiid phylogenetic dating analysis based on multiple genes and fossil and biogeographic calibration points (Burridge et al., 2012) suggests that the G. vulgaris radiation occurred largely within the Quaternary, especially when time-dependency of molecular rates is taken into account (Burridge et al., 2008b). This study estimated that G. depressiceps and G. ‘teviot’ diverged 0.60e1.98 mya [95% Highest Posterior Density]. Alternatively, divergence time of G. depressiceps and G. ‘teviot’ can be estimated by comparing levels of DNA sequence divergence or divergence time parameters relative to other galaxiid divergence events that are geologically constrained (Craw et al., 2008; Burridge et al., 2008b; Carrea et al., 2013). Net mtDNA sequence divergence between Taieri (G. depressiceps) and Teviot (G. ‘teviot’) lineages (which corrects for levels of divergence within lineages) is 2.73%. This degree of divergence can be compared with levels of net divergence for other known-date divergence events in galaxiid fishes (Craw et al., 2008), and suggests isolation of these catchments at a timeframe comparable to the Waiau-Oreti (140e240 ka) and Nevis-Mataura (300e500 ka) river divergences. Using the equation derived by Craw and coauthors (Craw et al., 2008) based on geologically-dated galaxiid divergence events, we estimate a divergence time of 276 ka for G. depressiceps and G. ‘teviot’. A similar period is suggested from the divergence time parameter derived under the four-parameter isolation model (Wakeley and Hey, 1997). Specifically, the divergence time parameter of 5.01 [2.55e17.01] is again comparable to that of the Waiau-Oreti divergence (Carrea et al., 2013). Overall, these molecular date estimates consistently support a mid-late Quaternary divergence event. 5. Geological estimates of age of capture Geological, geomorphic, and cosmogenic dating evidence suggests that the antiformal ranges in the vicinity of the present Taieri catchment have formed since the early Quaternary, with maximum detected ages >1 Ma (Bennett et al., 2005, 2006). Holocene uplift of

several metres along several km of laterally extending antiforms attest to rapid on-going crustal deformation. A 600 m high range with basement exposure age of 1.2 Ma (Bennett et al., 2006) implies an average uplift rate of ~0.5 mm per year (mm/a). Nearby coeval ranges are >1200 m high, implying local uplift rates exceeding 1 mm/a. Other cosmogenic data suggest uplift rates at the ends of developing ranges of ~0.1 mm/a, with lateral propagation of antiformal uplift occurring at ~1e8 mm/a (Bennett et al., 2005, 2006). As growing ranges meet and impinge on each other, the lateral propagation can evolve to enhanced vertical uplift and uplift rates, and this is a mechanism for ‘pinching off’, and reversal, of the intervening river (Jackson et al., 1996; Waters et al., 2001; Bennett et al., 2005, 2006). The current divide between the upper Taieri catchment and the Teviot River has resulted from this type of impingement of uplifting ranges, involving three different antiforms with underlying active fault structures (Figs. 1 and 3). Hence, uplift rates at the upper end of the estimates outlined above have probably been extant since the ranges started to physically interact. The lowest point of the present divide between Taieri and Teviot catchments is ~850 m above sea level, and is ~300 m above the almost-flat upper Taieri River (Fig. 3). Uplift of the basement barrier between these catchments will have taken 300 ka at an uplift rate of 1 mm/a, and 600 ka at an uplift rate of 0.5 mm/a. The higher uplift rate estimate agrees best with the estimates of divergence times for the galaxiids outlined in the previous section. However, it is also possible that genetic connection between the galaxiid populations continued after drainage reversal, via an initially-swampy divide region (wet connection (Burridge et al., 2006; Craw et al., 2007a)), so that the genetic divergence time is a minimum estimate for drainage reversal. Current data do not permit more precision with these timing estimates, but the galaxiid genetic data and geological estimates are broadly concordant in implying that drainage reversal occurred in the middle-late Quaternary (~300e600 ka). 6. Concluding remarks A combination of geological and biological evidence strongly supports a composite origin for the Taieri River. Specifically, we present evidence from structural geology (Figs. 1e3) and the presence of distinctive clasts in Quaternary fossil gravels (Fig. 3), to show that the ancestral Kye Burn originally drained south, through the Teviot valley, into the Clutha River, prior to its capture by the Taieri River during the mid-late Quaternary. Genetic dating analyses support the hypothesis that this transfer of a major river tributary between catchments led to divergence between formerly connected sister taxa G. depressiceps and G. ‘teviot’ during the midlate Quaternary. Fish distributional analyses lend further support to the geological hypothesis, as there is a substantial biogeographic disjunction between the lower- (ancestral) and upper (captured) portions of the Taieri River (Figs. 5 and 6). The robustness of this biogeographic inference is enhanced by the fact that geological and biological data are assessed independently (Bishop, 1995), yet yield consilient patterns and timeframes for the capture events inferred. Based on this study of Taieri drainage evolution, along with that of a previously documented river reversal involving a former tributary of a Southland river system (Nevis River (Waters et al., 2001)), we can conclude that the biological effects of major river reversal events can potentially be threefold: First, river capture can lead to increased biodiversity within a catchment. The geologically composite origin of the Taieri explains its unusually high galaxiid species diversity. A similar phenomenon has previously been noted for the Clutha system (Waters et al., 2001; Craw et al., 2008; Carrea et al., 2013). Second, transfer of tributaries among catchments can lead to vicariant isolation of formerly connected populations (Mayden,

J.M. Waters et al. / Quaternary Science Reviews 120 (2015) 47e56

1988; Bishop, 1995; Burridge et al., 2008b), and ultimately the evolution of new taxa. In the case of the Nevis, river reversal led to the vicariant formation of a distinctive endemic Nevis clade (Waters et al., 2001), currently attributed to Galaxias gollumoides but arguably now a distinct phylogenetic species (Waters et al., 2010). In the case of the Taieri, the formation of a new drainage divide within the ancestral Kye Burn (Fig. 2) led to the divergence of the similarly ‘stranded’ G. ‘teviot’ lineage (Fig. 4). Third, in addition to cladogenesis per se, major river-capture events can lead to biogeographic discontinuities within newly connected systems, such as the distributional disjunction between the upper and lower Taieri galaxiid faunas (Fig. 4), and between the Nevis and the rest of the Clutha (Waters et al., 2001). These apparently lingering associations between biological distributions/ relationships and historical (rather than contemporary) drainage patterns represent, in our view, an intriguing aspect of freshwater biogeography (Briggs et al., 1986; Mayden, 1988; Hurwood and Hughes, 1998; Near and Keck, 2005; Kozak et al., 2006). Although it could be argued that such patterns represent ‘the exception rather than the rule’, these distinctive associations seem to highlight an important role for historical (rather than contemporary ecological) factors in shaping biogeographic pattern (Ricklefs, 2004). We also suggest that such historic factors may sometimes be under-acknowledged in contemporary ecological biogeographic analyses (Leathwick et al., 2008). Understanding the retention of apparently ancient biogeographic patterns through to modern times requires a consideration of ecological factors that might affect species dispersal success and the structuring of biodiversity. In particular, explaining why lower Taieri (pullus; eldoni) and upper Taieri lineages (anomalus; depressiceps) have not infiltrated the ranges of one another since the Kye Burn capture event remains a challenge. Similarly, the reasons for the continued localised distributions of endemic Teviot (this study) and Nevis river lineages (Waters et al., 2001) are not immediately clear. One possibility, previously suggested by (Burridge et al., 2007), is that abiotic factors such as ecologically inhospitable gorges in the Clutha and Taieri river systems have prevented substantial range expansion through main-channel connections. An alternative ecological explanation is that ‘founder-takes-all’ demographic processes such as competitive exclusion and high-density blocking (previously discussed by (Waters, 2011; Waters et al., 2013)) have prevented substantial range-shifts since the capture event occurred (although we note that both G. anomalus and G. depressiceps have apparently expanded their ranges some 50 km south of the Kye Burn capture point, possibly outcompeting the lower Taieri lineages (Figs. 2 and 5)). It should be noted that these alternative explanations are not mutually exclusive. In favour of the latter explanation, however, phylogeographic studies in northern South Island have suggested that captured lineages can indeed undergo major range expansions when drainage changes translocate them into ecologically ‘vacant’ territory beyond their previous range limits (Burridge et al., 2006). Moreover, it seems likely that river capture followed by range expansion is the very mechanism responsible for the broad South Island distribution of the G. vulgaris complex as a whole (Fig. 4). In summary, this study highlights the interplay between physical and biological processes in a geologically dynamic setting. This type of interdisciplinary research approach is crucial to elucidating the processes governing the evolution and biogeography of freshwater biota. Additionally, the distributions of freshwater species can potentially elucidate the geometry of evolution of river drainages, and may provide key support for river capture events in cases where erosion has removed most or all the geological evidence (Cotterill and de Wit, 2011; Goodier et al., 2011).

55

Ethics statement Fish samples were collected under New Zealand Department of Conservation permits and multiple University of Otago Animal Ethics Permits. Authors' contributions JW, DC and GW conceptualised the research. JW analysed biogeographic and phylogenetic data, CB performed molecular dating analyses, and DC undertook geological analyses. All authors contributed to writing the manuscript. All authors gave final approval for publication. Funding statement Funding for this study was provided by Marsden Fund (Royal Society of NZ), University of Otago, and University of Tasmania. Acknowledgements The Department of Conservation assisted with sample collection, and Tania King assisted with DNA sequencing and analysis. Erica Todd provided helpful comments on an early version of the manuscript. References Allibone, R.M., Wallis, G.P., 1993. Genetic variation and diadromy in some New Zealand galaxiids (Teleostei: Galaxiidae). Biol. J. Linn. Soc. 50, 19e33. Allibone, R.M., Crowl, T.A., Holmes, J.M., King, T.M., McDowall, R.M., Townsend, C.R., Wallis, G.P., 1996. Isozyme analysis of Galaxias species (Teleostei: Galaxiidae) from the Taieri River, South Island, New Zealand: a species complex revealed. Biol. J. Linn. Soc. 57, 107e127. Apte, S., Smith, P.J., Wallis, G.P., 2007. Mitochondrial phylogeography of New Zealand freshwater crayfishes, Paranephrops spp. Mol. Ecol. 16, 1897e1908. Avise, J.C., 2000. Phylogeography. The History and Formation of Species. Harvard University Press, Cambridge, MA. Avise, J.C., Arnold, J., Ball, R.M., Bermingham, E., Lamb, T., Neigel, J.E., Reeb, C.A., Saunders, N.C., 1987. Intraspecific phylogeography: the mitochondrial DNA bridge between population genetics and systematics. Ann. Rev. Ecol. Syst. 18, 489e522. ~na ~rescu, P., 1990. Zoogeography of Fresh Waters. In: General Distribution and Ba Dispersal of Freshwater Animals, vol. 1. AULA-Verlag, Wiesbaden. Bennett, E.R., Youngson, J.H., Jackson, J.A., Norris, R.J., Raisbeck, G.M., Yiou, F., Fielding, E., 2005. Growth of South Rough Ridge, Central Otago, New Zealand: using in situ cosmogenic isotopes and geomorphology to study an active blind reverse fault. J. Geophys. Res. 110, B020404. http://dx.doi.org/10.1029/ 2004JB003184. Bennett, E.R., Youngson, J.H., Jackson, J.A., Norris, R.J., Raisbeck, G.M., Yiou, F., 2006. Combining geomorphic observations with in situ cosmogenic isotope measurements to study anticline growth and fault propagation in central Otago, New Zealand. N. Z. J. Geol. Geophys. 49, 217e231. Berra, T.M., 2001. Freshwater Fish Distribution. Academic Press, London. Bishop, P., 1995. Drainage rearrangement by river capture, beheading and diversion. Prog. Phys. Geog. 19, 449e473. Briggs, J.C., 1995. Global Biogeography. In: Developments in Palaeontology and Stratigraphy, 14. Elsevier, Amsterdam. Briggs, J.C., 1986. Introduction to the zoogeography of North American fishes. In: Hocutt, C.H., Wiley, E.O. (Eds.), The Zoogeography of North American Freshwater Fishes. John Wiley & Sons, New York, pp. 1e16. Burridge, C.P., Craw, D., Waters, J.M., 2006. River capture, range expansion, and cladogenesis: the genetic signature of freshwater vicariance. Evolution 60, 1038e1049. Burridge, C.P., Craw, D., Waters, J.M., 2007. An empirical test of freshwater vicariance via river capture. Mol. Ecol. 16, 1883e1895. Burridge, C.P., Craw, D., Jack, D.C., King, T.M., Waters, J.M., 2008a. Does fish ecology predict dispersal across a river drainage divide? Evolution 62, 1484e1499. Burridge, C.P., Craw, D., Fletcher, D., Waters, J.M., 2008b. Geological dates and molecular rates: fish DNA sheds light on time dependency. Mol. Biol. Evol. 25, 624e633. Burridge, C.P., McDowall, R.M., Craw, D., Wilson, M.V.H., Waters, J.M., 2012. Marine dispersal as a pre-requisite for Gondwanan vicariance among elements of the galaxiid fish fauna. J. Biogeogr. 39, 306e321. Carrea, C., Anderson, L.V., Craw, D., Waters, J.M., Burridge, C.P., 2013. The significance of past interdrainage connectivity for studies of diversity, distribution

56

J.M. Waters et al. / Quaternary Science Reviews 120 (2015) 47e56

and movement of freshwater-limited taxa within a catchment. J. Biogeogr. 41, 536e547. Chamberlain, C.P., Poage, M.A., Craw, D., Reynolds, R.C., 1999. Topographic development of the Southern Alps recorded by the isotopic composition of authigenic clay minerals, South Island, New Zealand. Chem. Geol. 155, 279e294. Coombs, D.S., Kawachi, Y., Landis, C.A., McDonough, W.F., Reay, A., 1986. Cenozoic volcanism in the north, east and central Otago. In: Smith, I.E.M. (Ed.), Late Cenozoic Volcanism in New Zealand, vol. 23. Royal Society of New Zealand, pp. 278e312. Cooper, A.F., Barriero, B.A., Kimbrough, D.L., Mattinson, D.L., 1987. Lamprophyre dike intrusion and the age of the Alpine Fault. Geology 15, 941e944. Cotterill, F.P.D., de Wit, M.J., 2011. Geoecodynamics and the Kalahari Epeirogeny: linking its genomic record, tree of life and palimpsest into a unified narrative of landscape evolution. S. Afr. J. Geol. 114, 489e514. Craw, D., Burridge, C., Anderson, L., Waters, J.M., 2007a. Late Quaternary river drainage and fish evolution, Southland, New Zealand. Geomorphology 84, 98e110. Craw, D., Burridge, C., Waters, J., 2007b. Geological and biological evidence for drainage reorientation during uplift of alluvial basins, central Otago, New Zealand. N. Z. J. Geol. Geophys. 50, 367e376. Craw, D., Burridge, C., Norris, R., Waters, J., 2008. Genetic ages for Quaternary topographic evolution: a new dating tool. Geology 36, 19e22. Craw, D., Upton, P., Walcott, R., Burridge, C., Waters, J., 2012. Tectonic controls on the evolution of the Clutha River catchment, New Zealand. N. Z. J. Geol. Geophys. 55, 345e359, 44. DeWoody, J.A., Avise, J.C., 2000. Microsatellite variation in marine, freshwater and anadromous fishes compared with other animals. J. Fish. Biol. 56, 461e473. Douglas, B., 1986. Lignite resources of Central Otago. N.Z. Energy Research and Development Committee, University of Auckland, Auckland, New Zealand, p. 368. Publication P104. Esa, Y.B., Waters, J.M., Wallis, G.P., 2000. Introgressive hybridization between Galaxias depressiceps and Galaxias sp D (Teleostei: Galaxiidae) in Otago, New Zealand: secondary contact mediated by water races. Cons. Genet. 1, 329e339. Fleming, C.A., 1979. The Geological History of New Zealand and its Life. Auckland University Press, Auckland. Forsyth, P.J., 2001. Geology of the Waitaki Area. Institute of Geological & Nuclear Sciences 1:250 000 Geological Map 19. 1 sheet þ 64 pp. Institute of Geological & Nuclear Sciences Ltd, Lower Hutt, New Zealand. Funk, V.A., Wagner, W.L., 1995. Biogeographic patterns in the Hawaiian islands. In: Wagner, W.L., Funk, V.A. (Eds.), Hawaiian Biogeography: Evolution on a Hot Spot Archipelago. Smithsonian Institution Press, Washington, pp. 379e419. Gillespie, R.G., Baldwin, B.G., Waters, J.M., Fraser, C., Nikula, R., Roderick, G.K., 2012. Long-distance dispersal: a framework for hypothesis testing. Trends Ecol. Evol. 27, 47e56. Goldberg, J., Trewick, S.A., Paterson, A.M., 2008. Evolution of New Zealand's terrestrial fauna: a review of molecular evidence. Phil. Trans. R. Soc. B 363, 3319e3334. Goodier, S.A., Cotterill, F.P., O'Ryan, C., Skelton, P.H., de Wit, M.J., 2011. Cryptic diversity of African tigerfish (Genus Hydrocynus) reveals palaeogeographic signatures of linked Neogene geotectonic events. PloS One 6, e28775. Gottscho, A.D., 2015. Zoogeography of the San Andreas Fault system: Great Pacific Fracture Zones correspond with spatially concordant phylogeographic boundaries in western North America. Biol. Rev. (in press). Heenan, P.B., McGlone, M.S., 2012. Evolution of New Zealand alpine and openhabitat plant species during the late Cenozoic. N. Z. J. Ecol. 37, 105e113. Hewitt, G.M., 2000. The genetic legacy of Quaternary ice ages. Nature 405, 907e913. Hoernle, K., White, J.D.L., van den Bogaard, P., Hauff, F., Coombs, D.S., Werner, R., € nberg, D., Reay, A., Cooper, A.F., 2006. Cenozoic intraplate Timm, C., Garbe-Scho volcanism on New Zealand: upwelling induced by lithospheric removal. Earth Planet. Sci. Lett. 248, 335e352. Hurwood, D.A., Hughes, J.M., 1998. Phylogeography of the freshwater fish, Mogurnda adspersa, in streams of northeastern Queensland, Australia: evidence for altered drainage patterns. Mol. Ecol. 7, 1507e1517. Jackson, J., Norris, R.J., Youngson, J.H., 1996. The structural evolution of active fault and fold systems in Central Otago, New Zealand: evidence revealed by drainage patterns. J. Struct. Geol. 18, 217e234. Juan, C., Ibrahim, K.M., Oromi, P., Hewitt, G.M., 2000. Colonization and diversification: towards a phylogeographic synthesis for the Canary Islands. Trends Ecol. Evol. 15, 104e109. Kozak, K.H., Blaine, R.A., Larsen, A., 2006. Gene lineages and eastern North American palaeodrainage basins: phylogeography and speciation in salamanders of the Eurycea bislineata species complex. Mol. Ecol. 15, 191e207. Landis, C.A., Campbell, H.J., Begg, J.G., Mildenhall, D.C., Paterson, A.M., Trewick, S.A., 2008. The Waipounamu Erosion Surface: Questioning the antiquity of the New Zealand land surface and terrestrial fauna and flora. Geol. Mag. 145, 173e197.

Leathwick, J.R., Elith, J., Chadderton, W.L., Rowe, D., Hastie, T., 2008. Dispersal, disturbance and the contrasting biogeographies of New Zealand's diadromous and non-diadromous fish species. J. Biogeogr. 35, 1481e1497. Lemey, P., et al., 2009. The Phylogenetic Handbook: a Practical Approach to Phylogenetic Analysis and Hypothesis Testing, second ed. Cambridge University Press. Leprieur, F., Hickey, M.A., Arbuckle, C.J., Closs, G.P., Brosse, S., Townsend, C.R., 2006. Hydrological disturbance benefits a native fish. J. Appl. Ecol. 43, 930e939. Lomolino, M.V., Riddle, B.R., Brown, J.H., 2005. Biogeography, third ed. Sinauer Associates, Sunderland, MA. Marsaglia, K.M., Martin, C.E., Kautz, C.Q., Shapiro, S.A., Carter, L., 2011. Linking a late Miocene-Pliocene hiatus in the deep-sea Bounty Fan off South Island, New Zealand, to onshore tectonism and lacustrine sediment storage. Geosphere 7, 305e312. Matthews, W.J., 1998. Patterns in Freshwater Fish Ecology. Chapman and Hall, London. Mayden, R.L., 1988. Vicariance biogeography, parsimony, and evolution in North American freshwater fishes. Syst. Zool. 37, 329e355. McDowall, R.M., 1970. The galaxiid fishes of New Zealand. Bull. Mus. Comp. Zool. 139, 341e432. McDowall, R.M., 1990. New Zealand freshwater Fishes e a Natural History and Guide. Heinemann, Auckland. McDowall, R.M., 1997. Two further new species of Galaxias (Teleostei: Galaxiidae) from the Taieri River, southern New Zealand. J. R. Soc. N. Z. 27, 199e217. McDowall, R.M., Wallis, G.P., 1996. Description and redescription of Galaxias species (Teleostei: Galaxiidae) from Otago and Southland. J. R. Soc. N. Z. 26, 401e427. Near, T.J., Keck, B.P., 2005. Dispersal, vicariance, and timing of diversification in Nothonotus darters. Mol. Ecol. 14, 3485e3496. Picard, D., Sempere, T., Plantard, O., 2008. Direction and timing of uplift propagation in the Peruvian Andes deduced from molecular phylogenetics of highland biotaxa. Earth Planet. Sci. Lett. 271, 326e336. Ricklefs, R.E., 2004. A comprehensive framework for global patterns in biodiversity. Ecol. Lett. 7, 1e15. Schwarzer, J., Misof, B., Ifuta, S.N., Schliewen, U.K., 2011. Time and origin of cichlid colonization of the lower Congo rapids. PLoS One 6, e22380. Shaw, K.L., 1996. Sequential radiations and patterns of speciation in the Hawaiian cricket genus Laupala inferred from DNA sequences. Evolution 50, 237e255. Sutherland, R., 1995. The Australia-Pacific boundary and Cenozoic plate motions in the SW Pacific: some constraints from Geosat data. Tectonics 14, 819e831. Turnbull, I.M., 2000. Geology of the Wakatipu Area. Institute of Geological & Nuclear Sciences 1:250 000 Geological Map 18. 1 sheet þ 72p. Institute of Geological & Nuclear Sciences Ltd, Lower Hutt, New Zealand. Upton, P., Craw, D., Walcott, R., 2014. Far-field deformation resulting from rheologic differences interacting with tectonic stresses: an example from the Pacific/ Australian plate boundary in southern New Zealand. Geosciences 4, 93e113. Wakeley, J., Hey, J., 1997. Estimating ancestral population parameters. Genetics 145, 847e855. Wallis, G.P., Trewick, S.A., 2009. New Zealand phylogeography: evolution on a small continent. Mol. Ecol. 18, 3548e3580. Wallis, G.P., Judge, K.F., Bland, J., Waters, J.M., Berra, T.M., 2001. Genetic diversity in New Zealand Galaxias vulgaris sensu lato (Teleostei: Osmeriformes: Galaxiidae). J. Biogeogr. 28, 59e67. Ward, R.D., Woodwark, M., Skibinski, D.O.F., 1994. A comparison of genetic diversity levels in marine, freshwater, and anadromous fishes. J. Fish. Biol. 44, 213e232. Waters, J.M., 2011. Competitive exclusion: phylogeography's ‘elephant in the room’? Mol. Ecol. 20, 4388e4394. Waters, J.M., Craw, D., Youngson, J.H., Wallis, G.P., 2001. Genes meet geology: fish phylogeographic pattern reflects ancient, rather than modern, drainage connections. Evolution 55, 1844e1851. Waters, J.M., Wallis, G.P., 2001a. Cladogenesis and loss of the marine life history phase in freshwater galaxiid fishes. Evolution 55, 587e597. Waters, J.M., Wallis, G.P., 2001b. Mitochondrial DNA phylogenetics of the Galaxias vulgaris complex from South Island, New Zealand: rapid radiation of a species flock. J. Fish. Biol. 58, 1166e1180. Waters, J.M., Rowe, D.L., Burridge, C.P., Wallis, G.P., 2010. Gene trees versus species trees: reassessing life-history evolution in a freshwater fish radiation. Syst. Biol. 59, 504e517. Waters, J.M., Fraser, C.I., Hewitt, G.M., 2013. Founder takes all: density-dependent processes structure biodiversity. Trends Ecol. Evol. 28, 78e85. White, K., Davidson, G.R., Paquin, P., 2009. Hydrologic evolution of the Edwards Aquifer recharge zone (Balcones fault zone) as recorded in the DNA of eyeless Cicurina cave spiders, south-central Texas. Geology 37, 339e342. Youngson, J., Craw, D., Landis, C., Schmitt, K., 1998. Redefinition and interpretation of late Miocene-Pleistocene terrestrial stratigraphy, Central Otago, New Zealand. N. Z. J. Geol. Geophys. 41, 51e68.