- Email: [email protected]
Palaeoecological signatures of vegetation change induced by herbivory regime shifts on subantarctic Enderby Island
Quaternary Science Reviews 134 (2016) 51e58
Contents lists available at ScienceDirect
Quaternary Science Reviews journal homepage: www.elsevier.com/locate/quascirev
Palaeoecological signatures of vegetation change induced by herbivory regime shifts on subantarctic Enderby Island Jamie R. Wood a, *, Janet M. Wilmshurst a, b, Chris S.M. Turney c, Christopher J. Fogwill c a
Long-Term Ecology Lab, Landcare Research, PO Box 69040, Lincoln 7640, New Zealand School of Environment, The University of Auckland, Private Bag 92019, Auckland 1142, New Zealand c Climate Change Research Centre, School of Biological, Earth and Environmental Sciences, University of New South Wales, NSW 2052, Australia b
a r t i c l e i n f o
a b s t r a c t
Article history: Received 20 October 2015 Received in revised form 20 December 2015 Accepted 21 December 2015 Available online 13 January 2016
The stratigraphic relationships of palaeoecological proxies and use of changepoint analyses to determine the cause and effect relationships between past events has allowed a better understanding of the relative contributions of humans and environmental drivers to Late Quaternary extinctions and of their effects on terrestrial ecosystems. Few studies, however, have validated these approaches at localities where past interactions between vegetation communities and large herbivores are well-documented. Here, we use a peat core from subantarctic Enderby Island to present the first study tracing the spores of dung fungi alongside pollen at a site where the history of mammalian herbivore introductions (and subsequent eradication), and their effects on the vegetation, are precisely known. We find a strong connection between spore influx rates of the dung-fungus Sporormiella and PCA axis 1 of the pollen assemblages, suggesting that past vegetation change caused by herbivore introductions and eradications at the core site can be readily deduced from the palaeoecological record. The response of the vegetation community to the removal of herbivores was so rapid, however, that a difference in timing between changepoints relating to specific pollen taxa, the overall pollen community, and the decline of Sporormiella spores, could not be resolved in our record, despite a sampling resolution of <5 years. We suggest that further case-studies, spanning different vegetation and herbivore communities, are required to provide increased confidence in inferences drawn about cause-and-effect relationships using proxy changepoint offsets in palaeoecological records. © 2016 Elsevier Ltd. All rights reserved.
Keywords: Dung-fungi Herbivores Holocene Palynology Palaeoecology Proxies Subantarctic
1. Introduction The past 50,000 years has been a period of major environmental change across the globe (Barnosky et al., 2004; Burney and Flannery, 2005). Human dispersal, climate change, shifts in vegetation communities, altered fire frequencies and major faunal extinction events have all had (though not necessarily independently) significant consequences for global ecosystems and ecological processes over a geologically short time frame (Janzen and Martin, 1982; Barlow, 2002; Johnson, 2009; Levy, 2012; Cooper et al., 2015). By determining the cause and effect relationships between these events we can gain a better understanding of
* Corresponding author. E-mail addresses: [email protected] (J.R. Wood), wilmshurstj@ landcareresearch.co.nz (J.M. Wilmshurst), [email protected] (C.S.M. Turney), [email protected] (C.J. Fogwill). http://dx.doi.org/10.1016/j.quascirev.2015.12.018 0277-3791/© 2016 Elsevier Ltd. All rights reserved.
the relative contributions of humans and environmental drivers to Late Quaternary extinctions of megafauna (Barnosky et al., 2004; Cooper et al., 2015; Bartlett et al., in press), examine the effects that these extinctions had on terrestrial ecosystems, and better predict the potential consequences of future environmental change and re-wilding projects. One way in which the cause and effect relationships between such past events can be inferred is by examining the sequence of changes in sedimentary proxies for each event. For example, within a sediment core, pollen assemblages (a proxy for vegetation community composition) can be stratigraphically connected with any number of other proxies, such as charcoal (a proxy for fire events), spores of dung-fungi (a proxy for the local density of large herbivores), or DNA of Bacterioides strain HF 183 (a proxy for the presence of local human populations) (Madeja et al., 2009, Madeja, 2015). Statistical methods for the detection of changepoints can then be used to determine when the trajectory of each proxy changes significantly, thereby allowing the relative timing of
52
J.R. Wood et al. / Quaternary Science Reviews 134 (2016) 51e58
changes between proxies to be inferred (e.g. Umbanhowar Jr. et al., 2006; Gill et al., 2012; Johnson et al., 2015). Recently, the relative stratigraphic order of changes in palaeoecological proxies have been used to test whether vegetation changes apparently coincident with the extinction of large herbivores in North America and Australia actually occurred before (i.e. contributed to the extinctions; Stuart et al., 2004) or after (i.e. were a response to reduced herbivory; Gill et al., 2009; Rule et al., 2012; Johnson et al., 2015) the event. However, while the relationships between sedimentary dung-fungi spores and herbivores, and pollen and vegetation, have been clearly demonstrated, there are relatively few case-studies showing how observed historical interactions between vegetation communities and large herbivores are expressed in palaeoecological records. Dull (1999) attempted to do this by using radiometric dating to establish a chronology for a core, from which known ages of herbivore introductions/extirpations were identified within the stratigraphy and thereby compared against the pollen record. Unfortunately, due to the inherent uncertainties associated with age-depth models, this approach is less precise than the comparison of proxies in a stratigraphic sequence where leads/lags can be determined more clearly. In this paper we examine the spores of dung fungi alongside pollen assemblages in a peat core taken from a subantarctic island in the southwest Pacific where the recent history of large herbivore introductions, their subsequent eradication less than 125 years later, and their effects on the vegetation are well-documented. Using ordination and changepoint analyses of the first detailed and radiocarbon dated palaeoecological record from this island, we aim to test how accurately the palaeoecological record reflects the true sequence of events, and thereby attempt to validate the approaches now being widely applied in Late Quaternary paleoecology.
established in 1865. The extermination of rabbits and cattle from Enderby Island for conservation purposes was completed in 1993 (Torr, 2002). Just prior to their extermination, rabbit densities of >35 individuals per hectare were recorded around Sandy Bay (Torr, 2002), and the cattle population consisted of 40e50 individuals, most of which were bulls (Department of Conservation (1998)). Rabbits and cattle had major impacts on the bird and vegetation communities at Sandy Bay, through processes including soil disturbance, grazing and destabilisation of the dunes, and these have been detailed by several authors during the 20th Century (e.g. Cockayne, 1904; Aston, 1912; Godley, 1965; Taylor, 1971). Moreover, the extermination of these mammals in 1993 led to a release of herbivory pressure and saw rapid changes in the vegetation communities at Sandy Bay (Torr, 2002).
2. Study site and methods
2.3. Palynology
2.1. Study site
Standard volumes of peat (1 mL) were taken from the peat core and processed by washing in hot KOH for 10 min followed by a wash in HCl, acetolysis, flotation of pollen and spores using lithium polytungstate (specific gravity 2.2), staining with fuchsin-red, and mounting on glass microscope slides in glycerol jelly. Samples were spiked with spores of exotic Lycopodium to allow pollen and spore concentrations to be quantified. A minimum of 200 pollen grains were identified on each slide (except for three slides with low concentration where a minimum of 80 grains were identified). Identifications were based on a reference collection of pollen grains made from herbarium specimens held at Landcare Research, Lincoln, New Zealand, and using pollen and spore identification keys (Large and Braggins, 1991; Moar, 1993). Pollen nomenclature follows Moar et al. (2011). The pollen concentrations in the Sandy Bay peat core varied markedly between the lower and upper portions of the sequence (56 and 54 cm). A minimum of 433,816 grains/mL was recorded before this zone, and maximum of 187,198 grains/mL afterwards. However, changes in pollen influx were ruled out as a cause because the composition of pollen assemblages did not vary dramatically across this zone. Accordingly, the most likely explanation of this change in pollen concentration is a change in deposition rate. Therefore, we created separate age-depth models for the bottom (using radiocarbon dates at 80 cm, 74 cm, 70 cm, 69 cm and 67 cm depths) and top (using radiocarbon dates at 56 cm, 30 cm and 16 cm, constrained by a collection year of 2013) of the sequence. The final age-depth model was created by linear interpolation between the modelled 95.4% confidence ranges for each dated horizon. For charcoal analysis, standard volumes of peat (2.5 mL) were bleached and passed through 250 mm and 125 mm mesh sieves. The
Our peat core (Landcare Research core X13/81a) was collected from behind the dunes at sea level on Sandy Bay, Enderby Island (50 290 55.8700 S, 166 170 2.2600 E) in December 2013 (Fig. 1) using a hand operated D-section corer. Enderby Island (700 ha) is the northernmost island of the subantarctic Auckland Island archipelago. Although there is a detailed post-glacial history of vegetation change from the Auckland Islands (McGlone et al., 2000, 2010; McGlone, 2002; Wilmshurst et al., 2015), the only pollen identified from Enderby Island includes spot samples from lake silts sandwiched between two possible Pleistocene-era tills (Fleming et al., 1976; McGlone, 2002). A midden containing stone tools and animal bones provides evidence for pre-European Polynesian settlement at Sandy Bay, but this was likely only of short duration (Anderson, 2005). Enderby Island was unsettled when it was first seen by Europeans in 1806. A number of shipwrecks and unsuccessful attempts at settling the island during the 19th Century saw a range of different mammal species being introduced (Taylor, 1971). The chronology of the mammalian herbivore introductions and eradications on Enderby Island is well documented (Taylor, 1971). Species introduced to the island included rabbits (Oryctolagus cuniculus), pigs (Sus scrofa), sheep (Ovis aries), cattle (Bos taurus) and goats (Capra hircus). Although 660 sheep were introduced to the island between 1850 and 1851 (Dingwall, 2009), rabbits and cattle had the longest persistence of any mammalian herbivores introduced to the island. Cattle (80 individuals) were introduced in 1850 (Dingwall, 2009) (although these only lasted a few years, and a persistent population of cattle were not introduced until 1894; Taylor, 1971) and a significant population of rabbits was first
2.2. Age-depth model Subsamples of bulk peat (spanning 1 cm intervals) were taken from eight locations through the peat sequence and submitted to the Waikato Radiocarbon Dating Laboratory for accelerator mass spectrometry dating. Samples were prepared by washing in hot HCl, rinsing, then treated with multiple hot NaOH washes. The fraction that was insoluble in NaOH was again treated with hot HCl, filtered, rinsed and dried. We created age-depth models for the Sandy Bay core using the P_sequence function of OxCal 4.2 (Bronk Ramsey, 2008). All modelled age ranges are reported in calibrated years AD. Radiocarbon ages were calibrated using the SHCal 13 calibration dataset (Hogg et al., 2013). Modern radiocarbon ages (i.e., post 1950 AD) were calibrated in Calibomb (http://calib.qub.ac.uk/CALIBomb; accessed Sept. 2015); using the SHCAL 13 and SHZ1-2 bomb extension zone options.
J.R. Wood et al. / Quaternary Science Reviews 134 (2016) 51e58
53
Fig. 1. Enderby Island, showing (a) location within the Auckland Island group and with respect to the New Zealand mainland, (b) Enderby Island, and c) location where the sediment core X13/81a was taken at Sandy Bay (dashed line marks approximate boundary of woody vegetation around the dunes).
number of charcoal fragments was counted in the remains of each sieve size under a microscope.
concentrations (Fig. 2b), indicating that this period represents a genuine slowing of peat accumulation, rather than an erosional unconformity.
2.4. Data analyses 3.2. Palynology Stratigraphic pollen diagrams were plotted using the program ‘C2’ (version 1.7.4). Clustering of pollen assemblages into stratigraphic zones was performed using hierarchical clustering of BrayeCurtis Similarity distances computed using the package ‘ecodist’ in R (version 3.0.2). Differences in pollen assemblages throughout the core were assessed using the PCA function of the R. Bayesian changepoint analyses were performed using the ‘bcp’ package (version 1.1.5) for R with a MCMC chain length of 5000 and burn-in of 500. Changepoints were identified as those with posterior probabilities >50%. 3. Results 3.1. Age-depth model The oldest radiocarbon date obtained from the core was taken at 80 cm depth and returned a radiocarbon age of 650 ± 25 years (modelled age of 1310 AD - 1404 AD; Table 1). The age-depth model (Fig. 2a) reflects three distinct periods of varying deposition rates. At the base of the core (88-67 cm) the deposition rate is ca. 0.14 cm/ yr. This is followed by a phase (67-56 cm) where the deposition rate slows to ca. 0.02 cm/yr, after which it increases again to ca. 0.61 cm/ yr. The period of relatively low deposition rate (67-56 cm) corresponds to a section of the core with the highest pollen
The pollen assemblages for the Sandy Bay core (Fig. 3) comprise five distinct stratigraphic zones. Zone 1a (53e88 cm) is characterised by a relatively high abundance of Metrosideros umbellata and Myrsine divaricata pollen and monolete fern spores. Zone 1b (49e53 cm) represents a departure from this towards the composition of zone 2 (19e37 cm), which is characterised by a reduction of Metrosideros and Myrsine pollen and an increase in the abundance of Caryophyllaceae, Callitriche and Cyperaceae pollen types. Zone 1c (37e49 cm) represents a reversion back to the state of zone 1a. Zone 3 (0e19 cm) is characterised by a drop in Callitriche pollen abundance and a dominance of Poaceae and Cyperaceae pollen types. Spores from two taxa of dung-fungi (Sporormiella and Sordaria) were identified (see Plate II, 27 and 28 in van Geel and Aptroot, 2006). The highest abundances were observed in zones 1b and 2 (Fig. 3). As the abundances of dung-fungi spores can be influenced by vegetation changes if expressed as a percentage of the pollen sum (Wood and Wilmshurst, 2013) we also calculated the influx rates of the spores based using the age-depth model (Fig. 4). Influx rates of Sporormiella were highest between 19 and 53 cm depth, spanning zones 1b, 1c and 2. Based on the age-depth model for the core, this corresponds to the period spanning 1894e1961 to 1981e1997 AD (95.4% confidence intervals).
54
J.R. Wood et al. / Quaternary Science Reviews 134 (2016) 51e58
Table 1 Radiocarbon dates from core X13/81A of Sandy Bay, Enderby Island, and 95.4% confidence calibrated age ranges for each date based on the Oxcal P_sequence age-depth model. Depth (cm)
Lab number
Radiocarbon age
Error
95.4% Confidence calibrated age range (modelled)
16 30 56 67 69 70 74 80
Wk-41772 Wk-39071 Wk-41773 Wk-39613 Wk-39614 Wk-39072 Wk-39615 Wk-39616
111.7% modern (post-1950 AD) 124.6% modern (post-1950 AD) 95 592 555 558 514 650
±0.3 ±0.3 ±20 ±25 ±25 ±25 ±25 ±25
1988 1957 1886 1415 1411 1406 1401 1310
AD AD AD AD AD AD AD AD
e e e e e e e e
2000 1988 1958 1443 1436 1432 1427 1404
AD AD AD AD AD AD AD AD
Fig. 2. (A) Age depth model for core X13/81a from Sandy Bay, Enderby Island. Depths at which radiocarbon measurements were taken are shown as filled circles. The rapid increase in sedimentation rate between 60 and 55 cm is consistent with a major decrease in sediment pollen concentration (B).
3.3. Drivers of change and sequence of changepoints The PCA of pollen assemblages show that composition changed as a factor of depth (Fig. 5a). The section of the core in which most of the change along axis 1 takes place corresponds to that which also has the highest influx rates of Sporormiella (Fig. 5b). Changepoints in the Sporormiella influx and abundance curves were compared with the positions of changepoints in the overall pollen assemblage (PCA axis 1 score), and abundance curves of Callitriche, Poaceae and Cyperaceae pollen types (i.e. those pollen types that exhibited large changes in abundance between pollen zones) (Fig. 6). Different changepoints were recognised in the Sporormiella influx and abundance curves. Two of the changepoints that occurred in the abundance curve but not the influx curve corresponded to changepoints in the PCA axis 1 score curve, demonstrating how Sporormiella abundance (expressed as a percentage of the pollen sum) can be influenced by changes in the vegetation community (Wood and Wilmshurst, 2013; Baker et al., 2013). One changepoint (18 cm) is shared by each of the curves examined (Fig. 6) and relates to the decline of Sporormiella and Callitriche and increases in Cyperaceae and Poaceae. The date of this changepoint was modelled as 1984e1999 AD, and hence most likely reflects the extermination of herbivores on Enderby Island (1993 AD). This also represents the last major change in Callitriche pollen abundance and the first changepoint in Poaceae pollen abundance (Fig. 6). 4. Discussion 4.1. Reconciling palaeoecological and historic records Detailed
botanical
observations
of
the
composition
of
vegetation communities on Enderby Island, and changes in vegetation communities at Sandy Bay, span more than a century (Cockayne, 1904, 1909; Taylor, 1971; Johnson and Campbell, 1975; Torr, 2002), providing a historical context against which the palaeoecological record can be interpreted. The earliest records of the vegetation at Sandy Bay were made during visits in 1840. McCormick (1884), who landed in Sandy bay in 1840, noted that the dunes were covered in tussock grasses which reached hip-height in damp hollows, and that these were fringed by bushes and trees. Hooker (1847) also described the vegetation at Sandy Bay as it was in 1840, noting that Pratia arenaria could be found creeping over the dunes. The presence of this species in the dunes was confirmed by Cockayne (1909). Through examining plant communities restricted to cliffs on Enderby Island through the presence of herbivores, and vegetation communities on subantarctic islands with no introduced herbivores, Taylor (1971) suggested that the pre-European dune vegetation at Sandy Bay was most likely dominated by the tussock grasses Poa litorosa and Poa foliosa, and by the megaherb Anisotome latifolia. The pre-European zone of our pollen core has little evidence for such vegetation, instead being dominated by pollen/ spores of forest species such as Metrosideros umbellata and Coprosma spp., ground ferns, and Myrsine divaricata, which locally forms a dense scrub ecotone around the forest margins (Taylor, 1971). Pollen types relating to the natural dune vegetation, including Apiaceae (i.e. Anisotome latifolia), Poaceae, Callitriche and Cyperaceae are relatively uncommon, suggesting that at this time the forest and scrub may have been much closer to the core site than it is today. A post-European recession of the forest margin at Sandy Bay is supported by observations of Cockayne (1904, 1909), who noted that disturbance by large herbivores had destabilised the dunes,
Fig. 3. Pollen diagram for core X13/81a from Sandy Bay, Enderby Island. Exaggeration lines are 10 .
J.R. Wood et al. / Quaternary Science Reviews 134 (2016) 51e58
55
which had begun to migrate over the margin of the Myrsine scrub and Metrosideros forest surrounding the bay. This recession of the local forest margin is reflected in the pollen diagram by a decrease in Coprosma and Myrsine pollen and ground fern (monolete) spores over time. Cockayne (1909) also noted that Rumex was a dominant plant on these mobile dunes, however, no Rumex pollen was recorded in our core. Although early European fires are noted to have played a part in altering forest margins on southern Enderby Island (e.g., a photograph from 1888 shows post-fire regeneration in an area behind Sandy Bay) (Taylor, 1971) we did not find charcoal evidence for these fires in our record (the only evidence for fire in our charcoal record likely relates to the burning of a hut at Sandy Bay in the 1990s). After the introduction of mammalian herbivores (in particular sheep, rabbits and cattle) to Enderby Island in the late 19th century the vegetation communities on the dunes at Sandy Bay began to change. Photographs of Sandy Bay, taken by Cockayne in 1903, show that tussock grasses were still growing around the edge of the woody vegetation that surrounded the dunes at this time (Taylor, 1971). The tussock grass Poa litorosa was observed at Sandy Bay by Aston (1912), but he noted that it appeared to be rapidly disappearing from cattle grazing. Davies (1919) makes little mention of tussock grasses, indicating that most of the tussocks may have been browsed out by 1919 (Taylor, 1971). Based on our age-depth model, the period when tussock grasses are reported to have been in decline corresponds to the increase in sedimentation rate occurring at ca. 56 cm depth. However, there is no concomitant loss of Poaceae pollen from the palaeoecological record, as native grasses had been a minor component of the pollen record prior to this time. On this point the historic observations and the pollen records are at odds, as McGlone and Moar (1997) found a close connection between the percentages of Poaceae in modern pollen assemblages vs. the percentage of Poaceae in the local vegetation. The intense grazing pressure that caused the decline of the tussock grasses on the dunes at Sandy Bay led to the replacement of native grasses with an exotic sward community (Fig. 7), comprised of a diverse assemblage of exotic grasses and small herbs (see Table 1 in Taylor, 1971). It appears that cattle may have been the key driver of this change, as on nearby Rose Island it was observed that tussocks reinvaded the sward following the eradication of cattle, which in turn led to the decline of the local rabbit population that favoured the sward habitat (Taylor, 1971). The turf is reflected in the pollen record by an increase in Callitriche, Stellaria and indeterminate Caryophyllaceae. Callitriche may not have been a dominant part of the vegetation community at Sandy Bay (Table 1 in Taylor, 1971; and G. Rogers pers comm.), however, it was found to be a major component of rabbit diets at that locality (Taylor, 1971). The high abundance of Callitriche pollen in the core may be due to this plant having relatively high fecundity. There is also an increase in Histiopteris spores in this zone. Taylor (1971) observed that Histiopteris was common in the forest understorey on Enderby Island. The increase of this fern in our pollen diagram suggests its abundance may have been promoted by disturbance as a result of herbivory. Pollen types of several ‘palatable’ taxa, such as the megaherbs Anisotome latifolia and Pleurophyllum and the small trees Raukaua simplex and Coprosma decline during this time. High population densities of herbivores are reflected in the high abundance and influx rates of dung-fungi spores. After the extermination of cattle and rabbits was completed on Enderby Island in 1993, there was a rapid response in the vegetation communities at Sandy Bay. These included a significant recovery of palatable plants (that had previously been restricted to inaccessible sites, e.g. Stilbocarpa polaris, Anisotome latifolia and Pleurophyllum criniferum) and the invasion of the sward by Poa litorosa (Torr, 2002). The extermination of herbivores coincides
56
J.R. Wood et al. / Quaternary Science Reviews 134 (2016) 51e58
Fig. 4. Sporormiella spore influx rate for (A) entire core X13/81a from Sandy Bay, Enderby Island, with (B) magnified section from 1800AD to present. The filled curve reflects the 95.4% confidence range of the age-depth model. Shown on the same time scale as B are the (C) known periods during which different introduced mammalian herbivores were present on Enderby Island.
Fig. 5. PCA plots of pollen assemblages in core X13/81a from Sandy Bay, Enderby Island. Assemblages are coloured by (A) stratigraphic depth and (B) corresponding Sporormiella spore influx rate at that depth. Grey lines connect stratigraphically contiguous assemblages.
with a major decline in the abundance and influx of dung-fungi spores, and a dramatic replacement of Callitriche pollen with Poaceae and Cyperaceae pollen types. The Cyperaceae pollen is morphologically attributable to Isolepis-type (Moar and Wilmshurst, 2003), and so may be from Isolepis aucklandicus, which was found to be common in a sample of turf from spongy peat on Enderby Island (Taylor, 1971). However, other sedges also grow on damp areas of Enderby Island (Cockayne, 1909). Therefore, the increase in Cyperaceae pollen (which begins with the
introduction of herbivores) may relate to increased wetness through cattle pugging and modification of water drainage in the dunes. The initial increase in Cyperaceae pollen in the core corresponds to an increase in deposition rate, which would be expected for peat growth in association with sedges. Grasses, particularly the tussocks Poa litorosa and Poa foliosa, were a common component of the dune vegetation before herbivores but were never a major component of the pollen assemblage. Pollen production varies greatly between different Poaceae species (Prieto-Baena et al.,
J.R. Wood et al. / Quaternary Science Reviews 134 (2016) 51e58
57
Fig. 6. Change-points identified in curves for Sporormiella, total pollen assemblage and key selected pollen types, plotted against the midpoint of the 95.4% confidence range of dates from the age-depth model.
Fig. 7. Photographs of dune vegetation at Sandy Bay, Enderby Island: A) view of dunes adjacent to beach ca. November 1907, (from the album: 1907 Sub-Antarctic Expedition, by Samuel Page. Te Papa (O.007033)); B) cattle grazing turfs at Sandy Bay, January 1966 (image from Taylor, 1971); C) view of dunes adjacent to beach in December 2013, taken near to the spot at which photo A was taken. Note the Poa tussock grasses have been replaced by turf vegetation; D) vegetation immediately surrounding the core site (marked by white circle).
2003), so, rather than reflecting a novel increase of grasses to a higher abundance than during the pre-human period, the postherbivore rise in grass pollen may have been driven by the return of different, more fecund exotic grass species. 4.2. Inferring cause and effect from paleoecological changepoints Our age-depth model provides a high level of agreement between the zone of elevated dung fungi spore abundance/influx and the time at which sustained herbivore populations were present on Enderby Island (Fig. 4). As detailed above, evidence from historic observations show that herbivores drove significant changes to the vegetation community at Sandy Bay. The connection between Sporormiella spore influx rates and the PCA of our pollen assemblages (Fig. 5b) shows that this change is reflected in the palaeoecological record. Therefore, it is clear that past vegetation change caused by herbivore introductions and eradications can be deduced from pollen cores. Our changepoint analysis showed that although there is a good
correspondence between several changepoints in the Sporormiella curves and the pollen/PCA curves, the final changepoint in Sporormiella curves is in perfect alignment with other proxies (Fig. 6). This observation suggests the vegetation response was so rapid that there was no lag in the paleoecological record that might allow a cause and effect relationship to be inferred were it not already known; this is despite the relatively high resolution sampling in this part of the core (sampled every 2 cm, equating to ca. 3.3 years resolution between samples). Our findings appears to suggest that caution may be required when interpreting changepoint offsets in Late Quaternary palaeoecological records, where resolution between samples may typically be on the order of decades or centuries. It is important to note, however, that our study represents just a single locality and vegetation type. Different vegetation communities may respond to herbivore introductions/extinctions at different rates (e.g. turf communities may respond much faster than tall forests), and may also respond differently to different herbivore community composition and densities.
58
J.R. Wood et al. / Quaternary Science Reviews 134 (2016) 51e58
5. Conclusions Our analyses of a peat core from Sandy Bay, Enderby Island, shows high congruence between dung-fungi spores (especially Sporormiella) and the known history of mammalian herbivores on the island, supporting the use of dung-fungi spores as a useful paleoecological proxy for herbivores. Moreover, changes in the local vegetation community that were a direct result of herbivore introductions are clearly resolvable in the pollen record. However, the response of vegetation communities to the extermination of these herbivores was so rapid that we could not resolve a timing difference between the decline of Sporormiella spores and changes in the pollen assemblage. As different vegetation communities may respond to herbivore introductions/extinctions at different rates, we suggest there is a need for further studies to better understand how known interactions between vegetation communities and large herbivores are expressed in palaeoecological records. Acknowledgements We thank the Department of Conservation for their support (DOC research permit National Authorisation # 37687-FAU) and the Australasian Antarctic Expedition 2013e2014 (www. spiritofmawson.com). We also thank B. Beaven, J. McDiarmid and V. Meduna for assistance in the field; K. Boot for laboratory sampling and macroscopic charcoal counts; and N. Bolstridge for preparing pollen slides. Our study was supported by Core Funding for Crown Research Institutes, from the New Zealand Ministry of Business, Innovation and Employment's Science and Innovation Group. References Anderson, A., 2005. Subpolar settlement in south Polynesia. Antiquity 79, 791e800. Aston, B.C., 1912. Some effects of imported animals on the indigenous vegetation. Trans. N. Z. Inst. 44, 19e24. Baker, A.G., Bhagwat, S.A., Willis, K.J., 2013. Do dung fungal spores make a good proxy for past distribution of large herbivores? Quat. Sci. Rev. 62, 21e31. Barlow, C., 2002. The Ghosts of Evolution: Nonsensical Fruit, Missing Partners, and Other Ecological Anachronisms. Basic Books, New York. Barnosky, A.D., Koch, P.L., Feranec, R.S., Wing, S.L., Shabel, A.B., 2004. Assessing the causes of late Pleistocene extinctions on the continents. Science 306, 70e75. Bartlett, L.J., Williams, D.R., Prescott, G.W., Balmford, A., Green, R.E., Eriksson, A., Valdes, P.J., Singarayer, J.S., Manica, A., 2015. Robustness despite uncertainty: regional climate data reveal the dominant role of humans in explaining global extinctions of late quaternary megafauna. Ecography. http://dx.doi.org/10.1111/ ecog.01566 (in press). Bronk Ramsey, C., 2008. Deposition models for chronological records. Quat. Sci. Rev. 27, 42e60. Burney, D.A., Flannery, T.F., 2005. Fifty millennia of catastrophic extinctions after human contact. Trends Ecol. Evol. 20, 395e401. Cockayne, L., 1904. A botanical excursion during midwinter to the southern islands of New Zealand. Trans. Proc. N. Z. Inst. 36, 225e333. Cockayne, L., 1909. The ecological botany of the subantarctic islands of New Zealand. In: Chilton, C. (Ed.), The Subantarctic Islands of New Zealand. Philosophical Institute of Canterbury, Wellington, NZ, pp. 182e235. Cooper, A., Turney, C., Hughen, K.A., Brook, B.W., McDonald, H.G., Bradshaw, C.J.A., 2015. Abrupt warming events drove Late Pleistocene Holarctic megafaunal turnover. Science 349, 602e606. Davies, J.K., 1919. With the 'Aurora' in the Antarctic, 1911-1914. Andrew Melrose, London. Department of Conservation, 1998. Conservation Management Strategy: Subantarctic Islands, 19982008. In: Southland Conservancy Conservation Management Planning Series, vol. 10. Department of Conservation, New Zealand. Dingwall, P.R., 2009. Pastoral farming at the Auckland Islands. In: Dingwall, P.R., Jones, K.L., Egerton, R. (Eds.), In Care of the Southern Ocean: an Archaeological and Historical Survey of the Auckland Islands. New Zealand Archaeological Association, Auckland, pp. 105e122. Dull, R.A., 1999. Palynological evidence for 19th century grazing-induced vegetation change in the southern Sierra Nevada, California, U.S.A. J. Biogeogr. 26,
899e912. Fleming, C.A., Mildenhall, D.C., Moar, N.T., 1976. Quaternary sediments and plant microfossils from Enderby Island, Auckland Islands. J. R. Soc. N. Z. 6, 433e458. Gill, J.L., Williams, J.W., Jackson, S.T., Lininger, K.B., Robinson, G.S., 2009. Pleistocene megafaunal collapse, novel plant communities, and enhanced fire regimes in North America. Science 326, 1100e1103. Gill, J.L., Williams, J.W., Jackson, S.T., Donnelly, J.P., Schellinger, G.C., 2012. Climatic and megaherbivory controls on late-glacial vegetation dynamics: a new, highresolution, multi-proxy record from Silver Lake, Ohio. Quat. Sci. Rev. 34, 66e80. Godley, E., 1965. Notes on the vegetation of the Auckland Islands. Proc. N. Z. Ecol. Soc. 12, 69e72. Hogg, A.G., Hua, Q., Blackwell, P.G., Niu, M., Buck, C.E., Guilderson, T.P., Heaton, T.J., Palmer, J.G., Reimer, P.J., Reimer, R.W., Turney, C.S.M., Zimmerman, S.R.H., 2013. SHCal13 southern Hemisphere calibration, 0e50,000 Years cal BP. Radiocarbon 55, 1889e1903. Hooker, J.D., 1847. Flora Antarctica: Botany of the Antarctic Voyage, vol. 1. Reeve Brothers, London. Janzen, D.H., Martin, P.S., 1982. Neotropical anachronisms: the fruits the gomphotheres ate. Science 215, 19e27. Johnson, C.N., 2009. Ecological consequences of late quaternary extinctions of megafauna. Proc. R. Soc. Lond. B 276, 2509e2519. Johnson, C.N., Rule, S., Haberle, S.G., Turney, C.S.M., Kershaw, A.P., Brook, B.W., 2015. Using dung fungi to interpret decline and extinction of megaherbivores: problems and solutions. Quat. Sci. Rev. 110, 107e113. Johnson, P.N., Campbell, D.J., 1975. Vascular plants of the auckland islands. N. Z. J. Bot. 13, 665e720. Large, M.F., Braggins, J.E., 1991. Spore Atlas of New Zealand Ferns and Fern Allies. SIR Publishing, Wellington, New Zealand. Levy, S., 2012. Once and Future Giants: what Ice Age Extinctions Tell Us About the Fate of Earth's Largest Animals. Oxford University Press, UK. McCormick, R., 1884. Voyages of Discovery in the Arctic and Antarctic Seas, and Round the World, vol. 1. Sampson Low, London. McGlone, M.S., 2002. The late quaternary peat, vegetation and climate history of the Southern Oceanic Islands of New Zealand. Quat. Sci. Rev. 21, 683e707. McGlone, M., Moar, N., 1997. Pollen-vegetation relationships on the subantarctic Auckland Islands, New Zealand. Rev. Palaeobot. Palynol. 96, 317e338. McGlone, M.S., Wilmshurst, J.M., Wiser, S.K., 2000. Late-glacial and Holocene vegetation and climate change on Auckland Island, subantarctic New Zealand. Holocene 10, 719e728. McGlone, M.S., Turney, C.S.M., Wilmshurst, J.M., Renwick, Pahnke, K., 2010. Divergent trends in land and ocean temperature in the Southern Ocean over the past 18,000 years. Nat. Geosci. 3, 622e626. Madeja, J., 2015. A new tool to trace past human presence from lake sediments: the human-specific molecular marker Bacteroides strain HF 183. J. Quat. Sci. 30, 349e354. Madeja, J., Wacnik, A., Zyga, A., Stankiewicz, E., Wypasek, E., Guminski, W., Harmata, K., 2009. Bacterial ancient DNA as an indicator of human presence in the past: its correlation with palynological and archaeological data. J. Quat. Sci. 24, 317e321. Moar, N.T., 1993. Pollen Grains of New Zealand Dicotyledonous Plants. Manaaki Whenua Press, Lincoln, New Zealand. Moar, N.T., Wilmshurst, J.M., 2003. A key to the pollen of the New Zealand Cyperaceae. N. Z. J. Bot. 41, 325e334. Moar, N.T., Wilmshurst, J.M., McGlone, M.S., 2011. Standardising names applied to pollen and spores in Quaternary New Zealand palynology. N. Z. J. Bot. 49, 201e229. Prieto-Baena, J.C., Hidalgo, P.J., Dominguez, E., Galan, C., 2003. Pollen production in the Poaceae family. Grana 42, 153e160. Rule, S., Brook, B.W., Haberle, S.G., Turney, C.S.M., Kershaw, A.P., Johnson, C.N., 2012. The aftermath of megafaunal extinction: ecosystem transformation in Pleistocene Australia. Science 335, 1483e1486. Stuart, A.J., Kosintsev, P.A., Higham, T.F.G., Lister, A.M., 2004. Pleistocene to Holocene extinction dynamics in giant deer and woolly mammoth. Nature 431, 684e689. Taylor, R.H., 1971. Influence of man on vegetation and wildlife of Enderby and Rose Islands, Auckland Islands. N. Z. J. Bot. 9, 225e268. Torr, N., 2002. Eradication of rabbits and mice from subantarctic Enderby and Rose Islands. In: Veitch, C.R., Clout, M.N. (Eds.), Turning the Tide: the Eradication of Invasive Species. IUCN SSC Invasive Species Specialist Group, IUCN, Gland, Switzerland and Cambrdige, UK, pp. 319e328. Umbanhowar Jr., C.E., Camill, P., Geiss, C.E., Teed, R., 2006. Asymmetric vegetation responses to mid-Holocene aridity at the prairieeforest ecotone in southcentral Minnesota. Quat. Res. 66, 53e66. van Geel, B., Aptroot, A., 2006. Fossil ascomycetes in quaternary deposits. Nova Hedwig. 82, 313e329. Wilmshurst, J.M., McGlone, M.S., Turney, C.S.M., 2015. Long-term ecology resolves the timing, region of origin, and process of establishment for a disputed alien tree. AOB Plants 7, plv104. Wood, J.R., Wilmshurst, J.M., 2013. Accumulation rates or percentages? How to quantify Sporormiella and other coprophilous fungal spores to detect late quaternary megafaunal extinctions. Quat. Sci. Rev. 77, 1e3.
Contents lists available at ScienceDirect
Quaternary Science Reviews journal homepage: www.elsevier.com/locate/quascirev
Palaeoecological signatures of vegetation change induced by herbivory regime shifts on subantarctic Enderby Island Jamie R. Wood a, *, Janet M. Wilmshurst a, b, Chris S.M. Turney c, Christopher J. Fogwill c a
Long-Term Ecology Lab, Landcare Research, PO Box 69040, Lincoln 7640, New Zealand School of Environment, The University of Auckland, Private Bag 92019, Auckland 1142, New Zealand c Climate Change Research Centre, School of Biological, Earth and Environmental Sciences, University of New South Wales, NSW 2052, Australia b
a r t i c l e i n f o
a b s t r a c t
Article history: Received 20 October 2015 Received in revised form 20 December 2015 Accepted 21 December 2015 Available online 13 January 2016
The stratigraphic relationships of palaeoecological proxies and use of changepoint analyses to determine the cause and effect relationships between past events has allowed a better understanding of the relative contributions of humans and environmental drivers to Late Quaternary extinctions and of their effects on terrestrial ecosystems. Few studies, however, have validated these approaches at localities where past interactions between vegetation communities and large herbivores are well-documented. Here, we use a peat core from subantarctic Enderby Island to present the first study tracing the spores of dung fungi alongside pollen at a site where the history of mammalian herbivore introductions (and subsequent eradication), and their effects on the vegetation, are precisely known. We find a strong connection between spore influx rates of the dung-fungus Sporormiella and PCA axis 1 of the pollen assemblages, suggesting that past vegetation change caused by herbivore introductions and eradications at the core site can be readily deduced from the palaeoecological record. The response of the vegetation community to the removal of herbivores was so rapid, however, that a difference in timing between changepoints relating to specific pollen taxa, the overall pollen community, and the decline of Sporormiella spores, could not be resolved in our record, despite a sampling resolution of <5 years. We suggest that further case-studies, spanning different vegetation and herbivore communities, are required to provide increased confidence in inferences drawn about cause-and-effect relationships using proxy changepoint offsets in palaeoecological records. © 2016 Elsevier Ltd. All rights reserved.
Keywords: Dung-fungi Herbivores Holocene Palynology Palaeoecology Proxies Subantarctic
1. Introduction The past 50,000 years has been a period of major environmental change across the globe (Barnosky et al., 2004; Burney and Flannery, 2005). Human dispersal, climate change, shifts in vegetation communities, altered fire frequencies and major faunal extinction events have all had (though not necessarily independently) significant consequences for global ecosystems and ecological processes over a geologically short time frame (Janzen and Martin, 1982; Barlow, 2002; Johnson, 2009; Levy, 2012; Cooper et al., 2015). By determining the cause and effect relationships between these events we can gain a better understanding of
* Corresponding author. E-mail addresses: [email protected] (J.R. Wood), wilmshurstj@ landcareresearch.co.nz (J.M. Wilmshurst), [email protected] (C.S.M. Turney), [email protected] (C.J. Fogwill). http://dx.doi.org/10.1016/j.quascirev.2015.12.018 0277-3791/© 2016 Elsevier Ltd. All rights reserved.
the relative contributions of humans and environmental drivers to Late Quaternary extinctions of megafauna (Barnosky et al., 2004; Cooper et al., 2015; Bartlett et al., in press), examine the effects that these extinctions had on terrestrial ecosystems, and better predict the potential consequences of future environmental change and re-wilding projects. One way in which the cause and effect relationships between such past events can be inferred is by examining the sequence of changes in sedimentary proxies for each event. For example, within a sediment core, pollen assemblages (a proxy for vegetation community composition) can be stratigraphically connected with any number of other proxies, such as charcoal (a proxy for fire events), spores of dung-fungi (a proxy for the local density of large herbivores), or DNA of Bacterioides strain HF 183 (a proxy for the presence of local human populations) (Madeja et al., 2009, Madeja, 2015). Statistical methods for the detection of changepoints can then be used to determine when the trajectory of each proxy changes significantly, thereby allowing the relative timing of
52
J.R. Wood et al. / Quaternary Science Reviews 134 (2016) 51e58
changes between proxies to be inferred (e.g. Umbanhowar Jr. et al., 2006; Gill et al., 2012; Johnson et al., 2015). Recently, the relative stratigraphic order of changes in palaeoecological proxies have been used to test whether vegetation changes apparently coincident with the extinction of large herbivores in North America and Australia actually occurred before (i.e. contributed to the extinctions; Stuart et al., 2004) or after (i.e. were a response to reduced herbivory; Gill et al., 2009; Rule et al., 2012; Johnson et al., 2015) the event. However, while the relationships between sedimentary dung-fungi spores and herbivores, and pollen and vegetation, have been clearly demonstrated, there are relatively few case-studies showing how observed historical interactions between vegetation communities and large herbivores are expressed in palaeoecological records. Dull (1999) attempted to do this by using radiometric dating to establish a chronology for a core, from which known ages of herbivore introductions/extirpations were identified within the stratigraphy and thereby compared against the pollen record. Unfortunately, due to the inherent uncertainties associated with age-depth models, this approach is less precise than the comparison of proxies in a stratigraphic sequence where leads/lags can be determined more clearly. In this paper we examine the spores of dung fungi alongside pollen assemblages in a peat core taken from a subantarctic island in the southwest Pacific where the recent history of large herbivore introductions, their subsequent eradication less than 125 years later, and their effects on the vegetation are well-documented. Using ordination and changepoint analyses of the first detailed and radiocarbon dated palaeoecological record from this island, we aim to test how accurately the palaeoecological record reflects the true sequence of events, and thereby attempt to validate the approaches now being widely applied in Late Quaternary paleoecology.
established in 1865. The extermination of rabbits and cattle from Enderby Island for conservation purposes was completed in 1993 (Torr, 2002). Just prior to their extermination, rabbit densities of >35 individuals per hectare were recorded around Sandy Bay (Torr, 2002), and the cattle population consisted of 40e50 individuals, most of which were bulls (Department of Conservation (1998)). Rabbits and cattle had major impacts on the bird and vegetation communities at Sandy Bay, through processes including soil disturbance, grazing and destabilisation of the dunes, and these have been detailed by several authors during the 20th Century (e.g. Cockayne, 1904; Aston, 1912; Godley, 1965; Taylor, 1971). Moreover, the extermination of these mammals in 1993 led to a release of herbivory pressure and saw rapid changes in the vegetation communities at Sandy Bay (Torr, 2002).
2. Study site and methods
2.3. Palynology
2.1. Study site
Standard volumes of peat (1 mL) were taken from the peat core and processed by washing in hot KOH for 10 min followed by a wash in HCl, acetolysis, flotation of pollen and spores using lithium polytungstate (specific gravity 2.2), staining with fuchsin-red, and mounting on glass microscope slides in glycerol jelly. Samples were spiked with spores of exotic Lycopodium to allow pollen and spore concentrations to be quantified. A minimum of 200 pollen grains were identified on each slide (except for three slides with low concentration where a minimum of 80 grains were identified). Identifications were based on a reference collection of pollen grains made from herbarium specimens held at Landcare Research, Lincoln, New Zealand, and using pollen and spore identification keys (Large and Braggins, 1991; Moar, 1993). Pollen nomenclature follows Moar et al. (2011). The pollen concentrations in the Sandy Bay peat core varied markedly between the lower and upper portions of the sequence (56 and 54 cm). A minimum of 433,816 grains/mL was recorded before this zone, and maximum of 187,198 grains/mL afterwards. However, changes in pollen influx were ruled out as a cause because the composition of pollen assemblages did not vary dramatically across this zone. Accordingly, the most likely explanation of this change in pollen concentration is a change in deposition rate. Therefore, we created separate age-depth models for the bottom (using radiocarbon dates at 80 cm, 74 cm, 70 cm, 69 cm and 67 cm depths) and top (using radiocarbon dates at 56 cm, 30 cm and 16 cm, constrained by a collection year of 2013) of the sequence. The final age-depth model was created by linear interpolation between the modelled 95.4% confidence ranges for each dated horizon. For charcoal analysis, standard volumes of peat (2.5 mL) were bleached and passed through 250 mm and 125 mm mesh sieves. The
Our peat core (Landcare Research core X13/81a) was collected from behind the dunes at sea level on Sandy Bay, Enderby Island (50 290 55.8700 S, 166 170 2.2600 E) in December 2013 (Fig. 1) using a hand operated D-section corer. Enderby Island (700 ha) is the northernmost island of the subantarctic Auckland Island archipelago. Although there is a detailed post-glacial history of vegetation change from the Auckland Islands (McGlone et al., 2000, 2010; McGlone, 2002; Wilmshurst et al., 2015), the only pollen identified from Enderby Island includes spot samples from lake silts sandwiched between two possible Pleistocene-era tills (Fleming et al., 1976; McGlone, 2002). A midden containing stone tools and animal bones provides evidence for pre-European Polynesian settlement at Sandy Bay, but this was likely only of short duration (Anderson, 2005). Enderby Island was unsettled when it was first seen by Europeans in 1806. A number of shipwrecks and unsuccessful attempts at settling the island during the 19th Century saw a range of different mammal species being introduced (Taylor, 1971). The chronology of the mammalian herbivore introductions and eradications on Enderby Island is well documented (Taylor, 1971). Species introduced to the island included rabbits (Oryctolagus cuniculus), pigs (Sus scrofa), sheep (Ovis aries), cattle (Bos taurus) and goats (Capra hircus). Although 660 sheep were introduced to the island between 1850 and 1851 (Dingwall, 2009), rabbits and cattle had the longest persistence of any mammalian herbivores introduced to the island. Cattle (80 individuals) were introduced in 1850 (Dingwall, 2009) (although these only lasted a few years, and a persistent population of cattle were not introduced until 1894; Taylor, 1971) and a significant population of rabbits was first
2.2. Age-depth model Subsamples of bulk peat (spanning 1 cm intervals) were taken from eight locations through the peat sequence and submitted to the Waikato Radiocarbon Dating Laboratory for accelerator mass spectrometry dating. Samples were prepared by washing in hot HCl, rinsing, then treated with multiple hot NaOH washes. The fraction that was insoluble in NaOH was again treated with hot HCl, filtered, rinsed and dried. We created age-depth models for the Sandy Bay core using the P_sequence function of OxCal 4.2 (Bronk Ramsey, 2008). All modelled age ranges are reported in calibrated years AD. Radiocarbon ages were calibrated using the SHCal 13 calibration dataset (Hogg et al., 2013). Modern radiocarbon ages (i.e., post 1950 AD) were calibrated in Calibomb (http://calib.qub.ac.uk/CALIBomb; accessed Sept. 2015); using the SHCAL 13 and SHZ1-2 bomb extension zone options.
J.R. Wood et al. / Quaternary Science Reviews 134 (2016) 51e58
53
Fig. 1. Enderby Island, showing (a) location within the Auckland Island group and with respect to the New Zealand mainland, (b) Enderby Island, and c) location where the sediment core X13/81a was taken at Sandy Bay (dashed line marks approximate boundary of woody vegetation around the dunes).
number of charcoal fragments was counted in the remains of each sieve size under a microscope.
concentrations (Fig. 2b), indicating that this period represents a genuine slowing of peat accumulation, rather than an erosional unconformity.
2.4. Data analyses 3.2. Palynology Stratigraphic pollen diagrams were plotted using the program ‘C2’ (version 1.7.4). Clustering of pollen assemblages into stratigraphic zones was performed using hierarchical clustering of BrayeCurtis Similarity distances computed using the package ‘ecodist’ in R (version 3.0.2). Differences in pollen assemblages throughout the core were assessed using the PCA function of the R. Bayesian changepoint analyses were performed using the ‘bcp’ package (version 1.1.5) for R with a MCMC chain length of 5000 and burn-in of 500. Changepoints were identified as those with posterior probabilities >50%. 3. Results 3.1. Age-depth model The oldest radiocarbon date obtained from the core was taken at 80 cm depth and returned a radiocarbon age of 650 ± 25 years (modelled age of 1310 AD - 1404 AD; Table 1). The age-depth model (Fig. 2a) reflects three distinct periods of varying deposition rates. At the base of the core (88-67 cm) the deposition rate is ca. 0.14 cm/ yr. This is followed by a phase (67-56 cm) where the deposition rate slows to ca. 0.02 cm/yr, after which it increases again to ca. 0.61 cm/ yr. The period of relatively low deposition rate (67-56 cm) corresponds to a section of the core with the highest pollen
The pollen assemblages for the Sandy Bay core (Fig. 3) comprise five distinct stratigraphic zones. Zone 1a (53e88 cm) is characterised by a relatively high abundance of Metrosideros umbellata and Myrsine divaricata pollen and monolete fern spores. Zone 1b (49e53 cm) represents a departure from this towards the composition of zone 2 (19e37 cm), which is characterised by a reduction of Metrosideros and Myrsine pollen and an increase in the abundance of Caryophyllaceae, Callitriche and Cyperaceae pollen types. Zone 1c (37e49 cm) represents a reversion back to the state of zone 1a. Zone 3 (0e19 cm) is characterised by a drop in Callitriche pollen abundance and a dominance of Poaceae and Cyperaceae pollen types. Spores from two taxa of dung-fungi (Sporormiella and Sordaria) were identified (see Plate II, 27 and 28 in van Geel and Aptroot, 2006). The highest abundances were observed in zones 1b and 2 (Fig. 3). As the abundances of dung-fungi spores can be influenced by vegetation changes if expressed as a percentage of the pollen sum (Wood and Wilmshurst, 2013) we also calculated the influx rates of the spores based using the age-depth model (Fig. 4). Influx rates of Sporormiella were highest between 19 and 53 cm depth, spanning zones 1b, 1c and 2. Based on the age-depth model for the core, this corresponds to the period spanning 1894e1961 to 1981e1997 AD (95.4% confidence intervals).
54
J.R. Wood et al. / Quaternary Science Reviews 134 (2016) 51e58
Table 1 Radiocarbon dates from core X13/81A of Sandy Bay, Enderby Island, and 95.4% confidence calibrated age ranges for each date based on the Oxcal P_sequence age-depth model. Depth (cm)
Lab number
Radiocarbon age
Error
95.4% Confidence calibrated age range (modelled)
16 30 56 67 69 70 74 80
Wk-41772 Wk-39071 Wk-41773 Wk-39613 Wk-39614 Wk-39072 Wk-39615 Wk-39616
111.7% modern (post-1950 AD) 124.6% modern (post-1950 AD) 95 592 555 558 514 650
±0.3 ±0.3 ±20 ±25 ±25 ±25 ±25 ±25
1988 1957 1886 1415 1411 1406 1401 1310
AD AD AD AD AD AD AD AD
e e e e e e e e
2000 1988 1958 1443 1436 1432 1427 1404
AD AD AD AD AD AD AD AD
Fig. 2. (A) Age depth model for core X13/81a from Sandy Bay, Enderby Island. Depths at which radiocarbon measurements were taken are shown as filled circles. The rapid increase in sedimentation rate between 60 and 55 cm is consistent with a major decrease in sediment pollen concentration (B).
3.3. Drivers of change and sequence of changepoints The PCA of pollen assemblages show that composition changed as a factor of depth (Fig. 5a). The section of the core in which most of the change along axis 1 takes place corresponds to that which also has the highest influx rates of Sporormiella (Fig. 5b). Changepoints in the Sporormiella influx and abundance curves were compared with the positions of changepoints in the overall pollen assemblage (PCA axis 1 score), and abundance curves of Callitriche, Poaceae and Cyperaceae pollen types (i.e. those pollen types that exhibited large changes in abundance between pollen zones) (Fig. 6). Different changepoints were recognised in the Sporormiella influx and abundance curves. Two of the changepoints that occurred in the abundance curve but not the influx curve corresponded to changepoints in the PCA axis 1 score curve, demonstrating how Sporormiella abundance (expressed as a percentage of the pollen sum) can be influenced by changes in the vegetation community (Wood and Wilmshurst, 2013; Baker et al., 2013). One changepoint (18 cm) is shared by each of the curves examined (Fig. 6) and relates to the decline of Sporormiella and Callitriche and increases in Cyperaceae and Poaceae. The date of this changepoint was modelled as 1984e1999 AD, and hence most likely reflects the extermination of herbivores on Enderby Island (1993 AD). This also represents the last major change in Callitriche pollen abundance and the first changepoint in Poaceae pollen abundance (Fig. 6). 4. Discussion 4.1. Reconciling palaeoecological and historic records Detailed
botanical
observations
of
the
composition
of
vegetation communities on Enderby Island, and changes in vegetation communities at Sandy Bay, span more than a century (Cockayne, 1904, 1909; Taylor, 1971; Johnson and Campbell, 1975; Torr, 2002), providing a historical context against which the palaeoecological record can be interpreted. The earliest records of the vegetation at Sandy Bay were made during visits in 1840. McCormick (1884), who landed in Sandy bay in 1840, noted that the dunes were covered in tussock grasses which reached hip-height in damp hollows, and that these were fringed by bushes and trees. Hooker (1847) also described the vegetation at Sandy Bay as it was in 1840, noting that Pratia arenaria could be found creeping over the dunes. The presence of this species in the dunes was confirmed by Cockayne (1909). Through examining plant communities restricted to cliffs on Enderby Island through the presence of herbivores, and vegetation communities on subantarctic islands with no introduced herbivores, Taylor (1971) suggested that the pre-European dune vegetation at Sandy Bay was most likely dominated by the tussock grasses Poa litorosa and Poa foliosa, and by the megaherb Anisotome latifolia. The pre-European zone of our pollen core has little evidence for such vegetation, instead being dominated by pollen/ spores of forest species such as Metrosideros umbellata and Coprosma spp., ground ferns, and Myrsine divaricata, which locally forms a dense scrub ecotone around the forest margins (Taylor, 1971). Pollen types relating to the natural dune vegetation, including Apiaceae (i.e. Anisotome latifolia), Poaceae, Callitriche and Cyperaceae are relatively uncommon, suggesting that at this time the forest and scrub may have been much closer to the core site than it is today. A post-European recession of the forest margin at Sandy Bay is supported by observations of Cockayne (1904, 1909), who noted that disturbance by large herbivores had destabilised the dunes,
Fig. 3. Pollen diagram for core X13/81a from Sandy Bay, Enderby Island. Exaggeration lines are 10 .
J.R. Wood et al. / Quaternary Science Reviews 134 (2016) 51e58
55
which had begun to migrate over the margin of the Myrsine scrub and Metrosideros forest surrounding the bay. This recession of the local forest margin is reflected in the pollen diagram by a decrease in Coprosma and Myrsine pollen and ground fern (monolete) spores over time. Cockayne (1909) also noted that Rumex was a dominant plant on these mobile dunes, however, no Rumex pollen was recorded in our core. Although early European fires are noted to have played a part in altering forest margins on southern Enderby Island (e.g., a photograph from 1888 shows post-fire regeneration in an area behind Sandy Bay) (Taylor, 1971) we did not find charcoal evidence for these fires in our record (the only evidence for fire in our charcoal record likely relates to the burning of a hut at Sandy Bay in the 1990s). After the introduction of mammalian herbivores (in particular sheep, rabbits and cattle) to Enderby Island in the late 19th century the vegetation communities on the dunes at Sandy Bay began to change. Photographs of Sandy Bay, taken by Cockayne in 1903, show that tussock grasses were still growing around the edge of the woody vegetation that surrounded the dunes at this time (Taylor, 1971). The tussock grass Poa litorosa was observed at Sandy Bay by Aston (1912), but he noted that it appeared to be rapidly disappearing from cattle grazing. Davies (1919) makes little mention of tussock grasses, indicating that most of the tussocks may have been browsed out by 1919 (Taylor, 1971). Based on our age-depth model, the period when tussock grasses are reported to have been in decline corresponds to the increase in sedimentation rate occurring at ca. 56 cm depth. However, there is no concomitant loss of Poaceae pollen from the palaeoecological record, as native grasses had been a minor component of the pollen record prior to this time. On this point the historic observations and the pollen records are at odds, as McGlone and Moar (1997) found a close connection between the percentages of Poaceae in modern pollen assemblages vs. the percentage of Poaceae in the local vegetation. The intense grazing pressure that caused the decline of the tussock grasses on the dunes at Sandy Bay led to the replacement of native grasses with an exotic sward community (Fig. 7), comprised of a diverse assemblage of exotic grasses and small herbs (see Table 1 in Taylor, 1971). It appears that cattle may have been the key driver of this change, as on nearby Rose Island it was observed that tussocks reinvaded the sward following the eradication of cattle, which in turn led to the decline of the local rabbit population that favoured the sward habitat (Taylor, 1971). The turf is reflected in the pollen record by an increase in Callitriche, Stellaria and indeterminate Caryophyllaceae. Callitriche may not have been a dominant part of the vegetation community at Sandy Bay (Table 1 in Taylor, 1971; and G. Rogers pers comm.), however, it was found to be a major component of rabbit diets at that locality (Taylor, 1971). The high abundance of Callitriche pollen in the core may be due to this plant having relatively high fecundity. There is also an increase in Histiopteris spores in this zone. Taylor (1971) observed that Histiopteris was common in the forest understorey on Enderby Island. The increase of this fern in our pollen diagram suggests its abundance may have been promoted by disturbance as a result of herbivory. Pollen types of several ‘palatable’ taxa, such as the megaherbs Anisotome latifolia and Pleurophyllum and the small trees Raukaua simplex and Coprosma decline during this time. High population densities of herbivores are reflected in the high abundance and influx rates of dung-fungi spores. After the extermination of cattle and rabbits was completed on Enderby Island in 1993, there was a rapid response in the vegetation communities at Sandy Bay. These included a significant recovery of palatable plants (that had previously been restricted to inaccessible sites, e.g. Stilbocarpa polaris, Anisotome latifolia and Pleurophyllum criniferum) and the invasion of the sward by Poa litorosa (Torr, 2002). The extermination of herbivores coincides
56
J.R. Wood et al. / Quaternary Science Reviews 134 (2016) 51e58
Fig. 4. Sporormiella spore influx rate for (A) entire core X13/81a from Sandy Bay, Enderby Island, with (B) magnified section from 1800AD to present. The filled curve reflects the 95.4% confidence range of the age-depth model. Shown on the same time scale as B are the (C) known periods during which different introduced mammalian herbivores were present on Enderby Island.
Fig. 5. PCA plots of pollen assemblages in core X13/81a from Sandy Bay, Enderby Island. Assemblages are coloured by (A) stratigraphic depth and (B) corresponding Sporormiella spore influx rate at that depth. Grey lines connect stratigraphically contiguous assemblages.
with a major decline in the abundance and influx of dung-fungi spores, and a dramatic replacement of Callitriche pollen with Poaceae and Cyperaceae pollen types. The Cyperaceae pollen is morphologically attributable to Isolepis-type (Moar and Wilmshurst, 2003), and so may be from Isolepis aucklandicus, which was found to be common in a sample of turf from spongy peat on Enderby Island (Taylor, 1971). However, other sedges also grow on damp areas of Enderby Island (Cockayne, 1909). Therefore, the increase in Cyperaceae pollen (which begins with the
introduction of herbivores) may relate to increased wetness through cattle pugging and modification of water drainage in the dunes. The initial increase in Cyperaceae pollen in the core corresponds to an increase in deposition rate, which would be expected for peat growth in association with sedges. Grasses, particularly the tussocks Poa litorosa and Poa foliosa, were a common component of the dune vegetation before herbivores but were never a major component of the pollen assemblage. Pollen production varies greatly between different Poaceae species (Prieto-Baena et al.,
J.R. Wood et al. / Quaternary Science Reviews 134 (2016) 51e58
57
Fig. 6. Change-points identified in curves for Sporormiella, total pollen assemblage and key selected pollen types, plotted against the midpoint of the 95.4% confidence range of dates from the age-depth model.
Fig. 7. Photographs of dune vegetation at Sandy Bay, Enderby Island: A) view of dunes adjacent to beach ca. November 1907, (from the album: 1907 Sub-Antarctic Expedition, by Samuel Page. Te Papa (O.007033)); B) cattle grazing turfs at Sandy Bay, January 1966 (image from Taylor, 1971); C) view of dunes adjacent to beach in December 2013, taken near to the spot at which photo A was taken. Note the Poa tussock grasses have been replaced by turf vegetation; D) vegetation immediately surrounding the core site (marked by white circle).
2003), so, rather than reflecting a novel increase of grasses to a higher abundance than during the pre-human period, the postherbivore rise in grass pollen may have been driven by the return of different, more fecund exotic grass species. 4.2. Inferring cause and effect from paleoecological changepoints Our age-depth model provides a high level of agreement between the zone of elevated dung fungi spore abundance/influx and the time at which sustained herbivore populations were present on Enderby Island (Fig. 4). As detailed above, evidence from historic observations show that herbivores drove significant changes to the vegetation community at Sandy Bay. The connection between Sporormiella spore influx rates and the PCA of our pollen assemblages (Fig. 5b) shows that this change is reflected in the palaeoecological record. Therefore, it is clear that past vegetation change caused by herbivore introductions and eradications can be deduced from pollen cores. Our changepoint analysis showed that although there is a good
correspondence between several changepoints in the Sporormiella curves and the pollen/PCA curves, the final changepoint in Sporormiella curves is in perfect alignment with other proxies (Fig. 6). This observation suggests the vegetation response was so rapid that there was no lag in the paleoecological record that might allow a cause and effect relationship to be inferred were it not already known; this is despite the relatively high resolution sampling in this part of the core (sampled every 2 cm, equating to ca. 3.3 years resolution between samples). Our findings appears to suggest that caution may be required when interpreting changepoint offsets in Late Quaternary palaeoecological records, where resolution between samples may typically be on the order of decades or centuries. It is important to note, however, that our study represents just a single locality and vegetation type. Different vegetation communities may respond to herbivore introductions/extinctions at different rates (e.g. turf communities may respond much faster than tall forests), and may also respond differently to different herbivore community composition and densities.
58
J.R. Wood et al. / Quaternary Science Reviews 134 (2016) 51e58
5. Conclusions Our analyses of a peat core from Sandy Bay, Enderby Island, shows high congruence between dung-fungi spores (especially Sporormiella) and the known history of mammalian herbivores on the island, supporting the use of dung-fungi spores as a useful paleoecological proxy for herbivores. Moreover, changes in the local vegetation community that were a direct result of herbivore introductions are clearly resolvable in the pollen record. However, the response of vegetation communities to the extermination of these herbivores was so rapid that we could not resolve a timing difference between the decline of Sporormiella spores and changes in the pollen assemblage. As different vegetation communities may respond to herbivore introductions/extinctions at different rates, we suggest there is a need for further studies to better understand how known interactions between vegetation communities and large herbivores are expressed in palaeoecological records. Acknowledgements We thank the Department of Conservation for their support (DOC research permit National Authorisation # 37687-FAU) and the Australasian Antarctic Expedition 2013e2014 (www. spiritofmawson.com). We also thank B. Beaven, J. McDiarmid and V. Meduna for assistance in the field; K. Boot for laboratory sampling and macroscopic charcoal counts; and N. Bolstridge for preparing pollen slides. Our study was supported by Core Funding for Crown Research Institutes, from the New Zealand Ministry of Business, Innovation and Employment's Science and Innovation Group. References Anderson, A., 2005. Subpolar settlement in south Polynesia. Antiquity 79, 791e800. Aston, B.C., 1912. Some effects of imported animals on the indigenous vegetation. Trans. N. Z. Inst. 44, 19e24. Baker, A.G., Bhagwat, S.A., Willis, K.J., 2013. Do dung fungal spores make a good proxy for past distribution of large herbivores? Quat. Sci. Rev. 62, 21e31. Barlow, C., 2002. The Ghosts of Evolution: Nonsensical Fruit, Missing Partners, and Other Ecological Anachronisms. Basic Books, New York. Barnosky, A.D., Koch, P.L., Feranec, R.S., Wing, S.L., Shabel, A.B., 2004. Assessing the causes of late Pleistocene extinctions on the continents. Science 306, 70e75. Bartlett, L.J., Williams, D.R., Prescott, G.W., Balmford, A., Green, R.E., Eriksson, A., Valdes, P.J., Singarayer, J.S., Manica, A., 2015. Robustness despite uncertainty: regional climate data reveal the dominant role of humans in explaining global extinctions of late quaternary megafauna. Ecography. http://dx.doi.org/10.1111/ ecog.01566 (in press). Bronk Ramsey, C., 2008. Deposition models for chronological records. Quat. Sci. Rev. 27, 42e60. Burney, D.A., Flannery, T.F., 2005. Fifty millennia of catastrophic extinctions after human contact. Trends Ecol. Evol. 20, 395e401. Cockayne, L., 1904. A botanical excursion during midwinter to the southern islands of New Zealand. Trans. Proc. N. Z. Inst. 36, 225e333. Cockayne, L., 1909. The ecological botany of the subantarctic islands of New Zealand. In: Chilton, C. (Ed.), The Subantarctic Islands of New Zealand. Philosophical Institute of Canterbury, Wellington, NZ, pp. 182e235. Cooper, A., Turney, C., Hughen, K.A., Brook, B.W., McDonald, H.G., Bradshaw, C.J.A., 2015. Abrupt warming events drove Late Pleistocene Holarctic megafaunal turnover. Science 349, 602e606. Davies, J.K., 1919. With the 'Aurora' in the Antarctic, 1911-1914. Andrew Melrose, London. Department of Conservation, 1998. Conservation Management Strategy: Subantarctic Islands, 19982008. In: Southland Conservancy Conservation Management Planning Series, vol. 10. Department of Conservation, New Zealand. Dingwall, P.R., 2009. Pastoral farming at the Auckland Islands. In: Dingwall, P.R., Jones, K.L., Egerton, R. (Eds.), In Care of the Southern Ocean: an Archaeological and Historical Survey of the Auckland Islands. New Zealand Archaeological Association, Auckland, pp. 105e122. Dull, R.A., 1999. Palynological evidence for 19th century grazing-induced vegetation change in the southern Sierra Nevada, California, U.S.A. J. Biogeogr. 26,
899e912. Fleming, C.A., Mildenhall, D.C., Moar, N.T., 1976. Quaternary sediments and plant microfossils from Enderby Island, Auckland Islands. J. R. Soc. N. Z. 6, 433e458. Gill, J.L., Williams, J.W., Jackson, S.T., Lininger, K.B., Robinson, G.S., 2009. Pleistocene megafaunal collapse, novel plant communities, and enhanced fire regimes in North America. Science 326, 1100e1103. Gill, J.L., Williams, J.W., Jackson, S.T., Donnelly, J.P., Schellinger, G.C., 2012. Climatic and megaherbivory controls on late-glacial vegetation dynamics: a new, highresolution, multi-proxy record from Silver Lake, Ohio. Quat. Sci. Rev. 34, 66e80. Godley, E., 1965. Notes on the vegetation of the Auckland Islands. Proc. N. Z. Ecol. Soc. 12, 69e72. Hogg, A.G., Hua, Q., Blackwell, P.G., Niu, M., Buck, C.E., Guilderson, T.P., Heaton, T.J., Palmer, J.G., Reimer, P.J., Reimer, R.W., Turney, C.S.M., Zimmerman, S.R.H., 2013. SHCal13 southern Hemisphere calibration, 0e50,000 Years cal BP. Radiocarbon 55, 1889e1903. Hooker, J.D., 1847. Flora Antarctica: Botany of the Antarctic Voyage, vol. 1. Reeve Brothers, London. Janzen, D.H., Martin, P.S., 1982. Neotropical anachronisms: the fruits the gomphotheres ate. Science 215, 19e27. Johnson, C.N., 2009. Ecological consequences of late quaternary extinctions of megafauna. Proc. R. Soc. Lond. B 276, 2509e2519. Johnson, C.N., Rule, S., Haberle, S.G., Turney, C.S.M., Kershaw, A.P., Brook, B.W., 2015. Using dung fungi to interpret decline and extinction of megaherbivores: problems and solutions. Quat. Sci. Rev. 110, 107e113. Johnson, P.N., Campbell, D.J., 1975. Vascular plants of the auckland islands. N. Z. J. Bot. 13, 665e720. Large, M.F., Braggins, J.E., 1991. Spore Atlas of New Zealand Ferns and Fern Allies. SIR Publishing, Wellington, New Zealand. Levy, S., 2012. Once and Future Giants: what Ice Age Extinctions Tell Us About the Fate of Earth's Largest Animals. Oxford University Press, UK. McCormick, R., 1884. Voyages of Discovery in the Arctic and Antarctic Seas, and Round the World, vol. 1. Sampson Low, London. McGlone, M.S., 2002. The late quaternary peat, vegetation and climate history of the Southern Oceanic Islands of New Zealand. Quat. Sci. Rev. 21, 683e707. McGlone, M., Moar, N., 1997. Pollen-vegetation relationships on the subantarctic Auckland Islands, New Zealand. Rev. Palaeobot. Palynol. 96, 317e338. McGlone, M.S., Wilmshurst, J.M., Wiser, S.K., 2000. Late-glacial and Holocene vegetation and climate change on Auckland Island, subantarctic New Zealand. Holocene 10, 719e728. McGlone, M.S., Turney, C.S.M., Wilmshurst, J.M., Renwick, Pahnke, K., 2010. Divergent trends in land and ocean temperature in the Southern Ocean over the past 18,000 years. Nat. Geosci. 3, 622e626. Madeja, J., 2015. A new tool to trace past human presence from lake sediments: the human-specific molecular marker Bacteroides strain HF 183. J. Quat. Sci. 30, 349e354. Madeja, J., Wacnik, A., Zyga, A., Stankiewicz, E., Wypasek, E., Guminski, W., Harmata, K., 2009. Bacterial ancient DNA as an indicator of human presence in the past: its correlation with palynological and archaeological data. J. Quat. Sci. 24, 317e321. Moar, N.T., 1993. Pollen Grains of New Zealand Dicotyledonous Plants. Manaaki Whenua Press, Lincoln, New Zealand. Moar, N.T., Wilmshurst, J.M., 2003. A key to the pollen of the New Zealand Cyperaceae. N. Z. J. Bot. 41, 325e334. Moar, N.T., Wilmshurst, J.M., McGlone, M.S., 2011. Standardising names applied to pollen and spores in Quaternary New Zealand palynology. N. Z. J. Bot. 49, 201e229. Prieto-Baena, J.C., Hidalgo, P.J., Dominguez, E., Galan, C., 2003. Pollen production in the Poaceae family. Grana 42, 153e160. Rule, S., Brook, B.W., Haberle, S.G., Turney, C.S.M., Kershaw, A.P., Johnson, C.N., 2012. The aftermath of megafaunal extinction: ecosystem transformation in Pleistocene Australia. Science 335, 1483e1486. Stuart, A.J., Kosintsev, P.A., Higham, T.F.G., Lister, A.M., 2004. Pleistocene to Holocene extinction dynamics in giant deer and woolly mammoth. Nature 431, 684e689. Taylor, R.H., 1971. Influence of man on vegetation and wildlife of Enderby and Rose Islands, Auckland Islands. N. Z. J. Bot. 9, 225e268. Torr, N., 2002. Eradication of rabbits and mice from subantarctic Enderby and Rose Islands. In: Veitch, C.R., Clout, M.N. (Eds.), Turning the Tide: the Eradication of Invasive Species. IUCN SSC Invasive Species Specialist Group, IUCN, Gland, Switzerland and Cambrdige, UK, pp. 319e328. Umbanhowar Jr., C.E., Camill, P., Geiss, C.E., Teed, R., 2006. Asymmetric vegetation responses to mid-Holocene aridity at the prairieeforest ecotone in southcentral Minnesota. Quat. Res. 66, 53e66. van Geel, B., Aptroot, A., 2006. Fossil ascomycetes in quaternary deposits. Nova Hedwig. 82, 313e329. Wilmshurst, J.M., McGlone, M.S., Turney, C.S.M., 2015. Long-term ecology resolves the timing, region of origin, and process of establishment for a disputed alien tree. AOB Plants 7, plv104. Wood, J.R., Wilmshurst, J.M., 2013. Accumulation rates or percentages? How to quantify Sporormiella and other coprophilous fungal spores to detect late quaternary megafaunal extinctions. Quat. Sci. Rev. 77, 1e3.