2.1. Typha biology, ecology, and morphology

The increasing distribution and abundance of Typha in wetland ecosystems around the world, particularly in North America, in recent decades is primarily attributed to anthropogenic-related disturbances to wetland hydrology and nutrient loads [11]. Typha is biologically and physiologically well-suited to take advantage of these disturbances (Fig 3). Typha produces an abundance of small, light (<100 μg), wind-dispersed seeds promoting colonisation of, and between, wetlands across great distances. The seeds germinate readily, within a few days under aquatic or high nutrient conditions [12]. But require a high light intensity for growth [13]. Ungerminated seeds may also remain viable in the wetland substrate for lengthy periods and form a persistent seed bank and source for expansion should suitable growth conditions develop.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3.

Schematic depictions of Typha individual (left) and community (right, reduced scale) illustrating various attributes promoting response to disturbance: 1. Large, resilient structures; 2. Small, wind-dispersed seeds; 3. Long seed-bank viability; 4. Flowers produce abundant pollen; 5. Aerated roots advantageous in organic-rich, flooded soils; 6. Rhizome carbohydrate storage maintains plant over winter and enables rapid spring growth; 7. Rhizomes anchor in substrate, trapping sediment and promoting lateral propagation; 8. Aggressive clonal propagation promotes dense monotypic stands; 9. Can spread across open water via floating rhizome mats.

https://doi.org/10.1371/journal.pwat.0000240.g003

Following seedling establishment, Typha develops axillary rhizomes with roots that anchor plants into wetland sediments. Rhizome tips develop into elongated linear leaves that grow rapidly and the plants develop a comparatively large stature which together with aggressive clonal propagation can result in dense monotypic stands [14–17]. A highly efficient root-aeration system adds to Typha’s competitive edge over other species, especially in organic-rich, flooded soils [18] and the plants can survive as floating mats which then opportunistically colonise newly disturbed sites. Typha rhizomes also store carbohydrates as the leaves grow and this serves to maintain the plant over its dormant winter period and to supply rapid new growth in spring. The dense mat of rhizomes retains debris resulting in a shallowing of the waterbody, enabling further rhizome propagation.

These biological attributes equip Typha to respond rapidly and aggressively to changing environmental conditions. Observations in North America, for example, report near pure Typha stands colonising an area of 40–80 m² within months to a few years [19, 20]. Disruption of natural hydrologic regimes and increased nutrient inputs into wetlands from agricultural runoff are repeatedly identified as underlying drivers of Typha invasion. Pollen studies have also shown that substantial increases in local abundances of Typha can rapidly follow landuse changes resulting in increased sediment and nutrient influxes to estuarine habitats [21]. Typha is also flood-tolerant and generally favoured by moderate flooding [22, 23]. Under naturally varying conditions, however, seasonal and inter-annual variability in water depths tends to keep Typha under control [2]. Conversely, a lack of variability, such as when water levels are controlled or artificially lowered, can reduce these seasonal extremes allowing the spread of Typha [24, 25]. In other situations, Typha spread can be enhanced by hydrologic alterations that raise the water table and create wetter soil conditions. In summary, Typha expansion into natural and controlled wetlands is often associated with a range of hydrologic alteration scenarios.

Sediment and water chemistry are also important factors in Typha growth and survival. Typha is often outcompeted by other aquatic macrophytes in low-nutrient, oligotrophic conditions [26, 27] but gains advantage under more eutrophic conditions. Increased nitrogen (N) and phosphorus (P) concentrations in wetland waters promote the growth of Typha, enabling its aggressive proliferation [26, 28, 29]. Typha invasion often follows increased sediment deposition in wetlands, due to enhanced inputs of sediment-attached N and P. Typha can also tolerate soils contaminated with heavy metals better than many of its competitors [5, 30, 31].

2.2. Benefits of Typha

2.2.1. Typha as a traditional source of food and materials.

Wherever Typha species occur in the world, they have been put to economic use by traditional societies [1], principally as sources of food and material serving a wide variety of functions. Although most of these observations relate to modern practice, in many cases they relate to long-established tradition practised within a particular culture and location across many generations, often extending into prehistoric times (Table 1). For example, Typha starch grains have been identified on the surface of a stone from the Paleolithic campsite of Bilancino (Florence, Italy), dated to around 25 000 BP, signifying that these and other local plants may have been used to grind into flour–presumably to make some kind of bread–by these ancient peoples long before the local ‘agricultural revolution’ [32]. In Aotearoa New Zealand, early European observations demonstrate that Māori use of raupō must also have prehistoric origins.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Traditional economic usage of Typha species around the world.

https://doi.org/10.1371/journal.pwat.0000240.t001

Most reported usages are of leaves and stems for making a variety of items ranging from footwear to huts and even houseboats (Table 1). In Aotearoa New Zealand, traditional lightweight boats for inland water transport were constructed from raupō stems (Fig 4). These craft, referred to as mōhiki (also, ‘mogi’), are especially common in the Waitaki and Clutha River regions (southern South Island/Te Waipounamu; Fig 2) where wānanga (workshops) are still held to teach this customary practice to younger generations [41, 42]. In some regions, entire dwellings, known as raupō whare (houses) were constructed from stems and leaves [40, 64].

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. A mōhiki constructed from raupō stems, Kurow museum (Aotearoa New Zealand). Photo: Rewi Newnham.

https://doi.org/10.1371/journal.pwat.0000240.g004

The soft floss or down from ripe Typha seeds, which are produced in abundance, have been employed in many regions as stuffing for cushions or pillows, as padding in cradles and quilts, and even for infant’s diapers (Table 1). In Australia, there are numerous observations from the early European era of Typha floss used for stuffing, particularly along the Murray River where it was sold under the name ‘Murray Down’ [36, 65].

Typha species have also been used worldwide as food [1]. Most prominent is the use of rhizomes and roots as a source of carbohydrates. Of particular relevance are the traditions practised by aboriginal Australians using the same species as in Aotearoa New Zealand (T. orientalis). Gott (1999) [36] reviews this widespread practice, which involved cooking the peeled rhizomes or roots either by steaming in an earth oven or roasting in fire ashes. Besides the rhizome, bases of mature shoots and leaves, the young shoots that appear in the spring and the raw young flower stalks are also eaten [58–60; all cited in 36].

In Aotearoa New Zealand, Crowe (2004) [57] reported that Māori also traditionally ate the peeled roots and the young shoots of raupō. More distinctive however was traditional Māori practice of making ‘cakes’ based on the copious raupō pollen. In 1880, the Reverend William Colenso gave an account of:

“…the pungapunga, the yellow pollen of the raupō flowers—the common bulrush… This was collected in the summer season, when the plant is in full flower, in the wet swamps and sides of lagoons, streams and lakes. I have been astonished at the large quantities of pollen then obtained. On one occasion, more than thirty years ago, I had several buckets full brought [to] me by the present chief, in his canoe, some of which I sent both raw and cooked to Kew Museum” (Bagnall and Petersen 1948, cited in Prendergast et al., 2000 [63]). These varying uses of raupō as food sources, while less prominent today, are still well documented and highlighted in many regional oral history accounts (e.g., T. Cassidy, personal communication, in: Šunde, 2022 [60]; N. Lomax and M. Heeney, personal communication, in: Šunde, 2022 [66]).

Similar usage of Typha pollen is reported from several other regions across the Northern Hemisphere (Table 1). Even today, Typha pollen sourced from the marshes situated at the confluences of the Tigris and Euphrates rivers is widely sold in the souks and cooperatives of Kuwait, having been mixed with sugar and steamed in a bag [63].

In summary, Māori have used raupō extensively as a food and material resource, in common with customary traditional practice throughout the temperate and tropical world. Its wide-ranging value as a resource raises questions as to the extent to which raupō was deliberately managed by pre-European Māori communities, including translocation by migrating tribes, and the extent of indigenous inter-generational knowledge and understanding acquired through these practices.

2.3 Previous Holocene records of raupō in Aotearoa New Zealand

Not surprisingly, published work from Aotearoa New Zealand fossil records shows patchy, discontinuous raupō presence in both space and time. Nevertheless, some interesting patterns stand out. First, consistent with its modern distribution, raupō is mostly absent from pollen records at sites above the lowland-montane zone (typically >800 m a.s.l) and tends to be more prominent at northern sites than in the south (Figs 2 and 5). For example, raupō pollen is present in late Holocene (the past several thousand years) sediments at Lake Coleridge, Canterbury (Fig 2; 510 m a.s.l) [83] but not recorded at all in three Holocene pollen and macrofossil records from the nearby Prospect Hill region (Fig 2) in the same catchment at 740–800 m asl [84].

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 5.

Reduced 2-taxa pollen diagrams showing raupō pollen and Pteridium spore percentages from lakes throughout Aotearoa New Zealand (A) Te Ika-a-Māui / North Island (B) Te Waipounamu / South Island. Inset: Chatham Islands. For each plot, Y axis is depth from surface of the lake bed in 50 cm increments and X axis increments are 25% for raupō and 50% for Pteridium. Blue triangles indicate earliest point after which sustained palynological evidence for human settlement occurred based on multiple palynological indicators including Pteridium (beginning of EMS). Areas in the maps coloured green) represent aggregated clusters of archaeological sites based on https://nzarchaeology.org/archsite. A strong, albeit variable, stratigraphic association is illustrated between raupō and Pteridium–and hence with human settlement patterns–throughout both main islands, within its presumed natural limits. Basemap showing Aotearoa New Zealand sourced from the LINZ Data Service and licensed for reuse under the CC BY 4.0 licence. (https://www.linz.govt.nz/products-services/data/licensing-and-using-data/attributing-linz-data).

https://doi.org/10.1371/journal.pwat.0000240.g005

Also consistent with modern observations, the raupō paleoecological record points to an opportunistic response to natural disturbance. In particular, where volcanic deposits are recorded at lower elevation pollen sites, raupō pollen often appears or increases prominently in the immediate overlying assemblages [9, 85–90]. Moreover, the plant’s affinity for nutrient-rich sites is apparent. Raupō pollen is rare from lowland acidic low nutrient Sphagnum [91] or restiaid peat bogs [92, 93] but can be prominent where nutrient levels are higher [9]. At a few sites, raupō pollen levels increase during phases of lake level lowering or hydroseral succession indicated by stratigraphic change [9].

All of the characteristics noted above are observed in Holocene pollen records from Lake Poukawa, eastern North Island/ Te Ika-a-Māui (Fig 2), a lowland (20 m a.s.l) shallow lake surrounded by extensive peat swamps that contained abundant raupō prior to drainage for pastoral agriculture in recent decades [88]. Pollen assemblages from drill sites taken beyond the current lake perimeter indicate raupō prominence during an earlier undated warm period, assumed to be the last interglacial (~125,000 years ago), and absence from sediments dated to the subsequent (last) glacial phase when presumably colder climates prevented its survival in this region. In contrast, Holocene lake sediments show highly variable raupō pollen percentages with peak levels during fen swamp phases reflecting lake level changes and immediately following tephra layers.

A final pattern, evident at some sites and consistent with modern observations, is that raupō often becomes more prominent in the anthropogenic era [90, 94–98]. This observation is not universal. For example, Wilmshurst (1997) [99] presents pollen diagrams from two lowland eastern North Island/ Te Ika-a-Māui sites that show no obvious raupō increase in the anthropogenic era, with continuous low pollen percentages being slightly higher overall in pre-anthropogenic era sediments at both sites. Caution is required in this interpretation. Whilst McGlone’s (2009) [9] review of wetlands in Aotearoa New Zealandnotes that raupō pollen is well represented and often completely dominates pollen sums, previous work, consistent with Typha pollen observations from North America [100], suggests it may not be well-dispersed beyond the lake margins where the plant grows [88, 101, 102]. Both lakes investigated by Wilmshurst (1997) [99] are comparatively deep and large enough in area for the pollen cores at the lakes’ depocentre to be insensitive to raupō variability in the littoral margins.

These observations from Holocene records point to overall raupō behaviour that is broadly consistent with modern ecological observations, but they are drawn from a relatively small number of records. The Lakes380 programme offers the opportunity to build upon these previous observations with a more targeted investigation of the natural and anthropogenic history of raupō.

3.1. Study lakes and sediment coring

The research programme Lakes380 (www.lakes380.com) aims to enrich understanding of the environmental, social, and cultural histories of lakes in Aotearoa New Zealand. These lakes cover a 12-degree latitudinal gradient and a range of environmental gradients including altitude, size, depth, trophic status, and geomorphic-catchment characteristics from coastal to alpine locations. Sediment cores from a total of 92 lake sites were analysed as part of this raupō study (S1 Table).

At each lake, four sediment cores were taken in close proximity at the deepest part of the lake using a UWITEC Gravity corer (Mondsee, Austria) fitted with two metre polyvinyl chloride (PVC; 90-mm dia.) core barrels. Upon extraction, core barrels were packed with florist foam to prevent sediment movement, and after settling, the barrels were cut into one metre lengths for transportation. Cores were kept refrigerated while in the field and in transit to the laboratory at GNS Science (National Isotope Centre, Lower Hutt, Aotearoa New Zealand), where the cores were split along the longitudinal plane, described, and imaged.

3.2. Palynology

Pollen was extracted from 0.25 cm³ sediment at a variable sampling interval of 1–2 and 3–4 cm in the upper 65 cm of the cores and up to 10 cm intervals for the lower section (>65 cm) of most cores. The variable sampling interval was employed to build a vegetation reconstruction with a focus on the boundaries of vegetation change that have occurred at these lake sites over the past ~1000 years with respect to pre- and post-human vegetation and landscape change. Pollen extraction was carried out following standard laboratory techniques [103], but methods were refined to streamline pollen processing based on the sedimentary characteristics of each lake with the goal to achieve good pollen recovery with minimum processing steps. Samples were prepared using 10% hot HCl, acetolysis, and 6 μm sieving. A density flotation was applied for strongly minerogenic sediments. Exotic Lycopodium tablets were added to each sample to allow the calculation of pollen concentrations. Pollen and spore identifications were made using standard texts [104–108] and Aotearoa New Zealand pollen reference collections at GNS Science. Pollen taxonomy follows Moar et al. (2011) [109].

Pollen data are presented in the form of relative frequency of a minimum pollen sum of 150 grains. This sum includes pollen from all dryland plants, i.e. trees, shrubs and herbaceous plants, non-native plant taxa, and in addition to 150 dryland sum the bracken fern Pteridium esculentum. P. esculentum is included in the dryland pollen sum as its growth form in a (post) disturbance landscape is closer in functional morphology to a shrub than a fern, and communities are ecologically equivalent to shrubland [110]. Pollen of other groups (wetland, aquatics, ferns, tree ferns as well as non-palynomorphs) were excluded from the pollen sum, but their percentages were calculated as a proportion of dryland pollen plus the respective group. It should be noted that interpreting relative frequency data at face value is problematic because of the compositional effect, but is preferable here to the use of alternative metrics such as pollen concentration or influx that are biased by site-specific, large and as yet unconstrained variations in sediment accumulation rates over the last 1 ka.

Charcoal was counted as number of fragments on the pollen slides and presented as concentration per cm³.

3.3. Principal component analysis

The primary objective of principal component analysis (PCA) in this study was to determine the stratigraphic relationships between raupō and other pollen taxa of interest, in particular in relation to the research hypothesis. PCA was applied to square-root transformed pollen percentages, scaled to unit variance. Only plant taxa that occur in at least 20 of the 92 sites (22%) and have a maximum abundance in the dataset of at least 2% were included. A list of included taxa can be found in the S1 Fig and S2 Table. PCA was performed in R v.4.1.0 [111] using package vegan [112]. Plots were created with package ggplot2 [113].

3.4. Chronology

Although radiometric dating has been undertaken for some of these records, this research primarily draws upon a well-established chronostratigraphic framework for the past millennium in Aotearoa New Zealand based on changes in dryland pollen [110, 114]. Throughout Aotearoa New Zealand, pollen records for this interval typically show no obvious evidence for human activity until ca. 1250 AD, consistent with material archaeological records [115]. From ca. 1250 AD, key indicators for early Māori activity, often referred to as the Polynesian era, are a marked rise of charcoal accompanying the sustained decline in tall tree pollen and an accompanying increase in disturbance indicators, typically Pteridium esculentum. As we cannot preclude non-visible human presence in the pollen records prior to these visible key indications of anthropogenic activity, we use the term Evidence for Māori Settlement (EMS) for this period. Prior to this period, we use the term pre-EMS, essentially a phase of natural variability. Finally, the European era (EE), commencing in the early 19^th century, is determined in our pollen records from the appearance of introduced plants typically associated with European settlement and agriculture such as Pinus and Rumex, and increased levels of Poaceae pollen reflecting pasture grasses.

4.1. Raupō distribution

From a total of 92 new pollen records, 32 records (35%) either have no observed raupō pollen (21 records; 23%) or trace amounts (<1%; 11 records; 12%) and most of these are typically located at higher elevations (>800 m asl) or in the far south of the South Island/Te Waipounamu (Fig 5B). Of the remaining 60 records, 46 records are ‘complete’ in that they span all three phases from pre-EMS to EE and the remaining 14 records sample only the EMS and EE phases (S2 Table). The 46 complete pollen records fall into three broad categories, depending upon the timing and stratigraphic pattern of raupō response. In addition, some of the 14 ‘incomplete’ records show similar characteristic stratigraphic patterns despite not encompassing the pre-EMS phase (S1 Table).

Category I (pre-EMS) is determined as those records where the first occurrence of raupō is detected during the pre-EMS phase, as seen in a small number of sites (13 sites; 22% of all sites), either at low amounts (% <5% raupō; 10 sites; 17% of all sites) or at a maximum (~17–40% raupō; 3 sites; 5% of all sites). At Lake Kereta, Kaipara Peninsula, Northland (Fig 6A), for example, the classic EMS indicators occur after a prominent peak in raupō along with sedges (Cyperaceae). Raupō and sedges then decline to background levels before both rise again, coincident with sustained increases in bracken, charcoal, and other indicators of the EMS era.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 6. Raupō and cyperaceae pollen and Pteridium spore curves along with summary ecological groups at key pollen sites indicating different temporal patterns of raupō increase (see text for description of categories).

Pollen and spore counts are percentages, charcoal counts are number of specimens per cm³ of sediment sample. (a) Category I, Lake Kereta; (b) Category II a, Lake Kohangapiripiri; (c) Category II b, Lake Mangarakau; (d) Category II c, Lake Te Kahika; (e) Category III, Lake Killarney. See S1 Table for information on lakes including location.

https://doi.org/10.1371/journal.pwat.0000240.g006

Category II (EMS) consists of the records where raupō is present and shows a clear pattern of response during the EMS era (35 sites; 58% of all sites). The precise timing and extent of response varies between sites. For example, at Lake Kohangapiripiri, near Wellington (Fig 6B), raupō rises for the first time in tandem with bracken and other anthropogenic disturbance indicators early in the EMS phase. At some other sites, raupō exhibits a distinctive bimodal pattern whereby it rises early in the EMS phase, then declines before rising again later in the phase, usually mirroring fluctuations in anthropogenic disturbance indicators. At Lake Mangarakau, northwestern South Island/Te Waipounamu (Fig 6C), the raupō pollen curve, along with other wetland indicators(in particular sedges) covaries positively with the charcoal curve as well as bracken, but negatively with some native forest indicators, such as Dacrydium cupressinum and Cyathea smithii. At other sites such as Lake Te Kahika in the Far North (Fig 6D), raupō rises to a peak late in the EMS phase, before declining gradually during the European era. At this site, and at most others in this category, other wetland or aquatic taxa show a similar stratigraphic pattern to raupō. At Lake Te Kahika, for example, sedges are first visible during the early EMS phase, but rise to maximum prominence late in the phase, in tandem with raupō.

Category III (EE) comprises those records (7 sites; 12% of all sites) where raupō is either not observed or is present at comparatively low levels until the European era (EE), when it rises to peak prominence. At Lake Killarney in northwest Nelson for example (Fig 6E), raupō pollen is absent from the record during the pre-EMS and EMS phases and is only visible in sediments dominated by introduced European plants. Collectively, of the 60 sites that contain raupō present at >1%, 87% of sites record raupō at its maximum palynological evidence post human arrival (during the EMS–EE eras), and only 5% of sites have raupō prominent during the pre-human era (S2 Table). The remaining five records (8%) are unclassified, where the pattern of raupō response is hard to discern due to the incompleteness of the records (S2 Table).

4.2. Principal components analysis

The PCA of the entire pollen dataset is presented in Fig 7, which displays the sample and prominent taxa scores on the first two principal components. Although the first two principal components represent comparatively low levels of explanation of variability in the dataset (13.1% and 8.0%; Fig 7), this is not unexpected given the high variability in vegetation composition across the diverse environment of Aotearoa New Zealand. Despite this variability, the 3-phase subdivision of the last ~1000 years is strongly represented by the sample scores along the first axis in particular, with pre-EMS sample scores mostly positive, EE mostly negative, and EMS scores intermediate between but also overlapping with the other two phases. It is interesting to note that this EMS overlap is stronger with the pre-EMS phase than with the EE phase.The taxa scores on the first PCA axis are consistent with this observation, with negative values for key pollen indicators for both human phases (EMS and EE) and positive values for those taxa most prominent during the pre-EMS phase. Notably, raupō clusters with the former grouping, indicating its strong stratigraphic association with human disturbance. From these observations we suggest that although raupō typically rises to prominence in the EMS phase, in response to disturbance, the disruption to the natural (pre-human) vegetation communities was less severe during the EMS phase than during the EE phase.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 7. Biplot of taxa and sample scores on the first two principal component (PC) axes for 86 pollen records from lakes in Aotearoa New Zealand, spanning approximately the last millennium.

The most common pollen taxa are labelled on the diagram. The same plot but with a more complete illustration of taxa scores is available in S1 Fig and S2 Table).

https://doi.org/10.1371/journal.pwat.0000240.g007

5.1. Raupō as a disturbance indicator

Our results indicate a consistent raupō response to disturbance, particularly accompanying catchment deforestation that characterises the comparatively brief human era in Aotearoa New Zealand. The strong negative PC1 score recorded for raupō pollen contrasts markedly with the positive scores recorded by native tree taxa, in particular the tall canopy and emergent podocarps (represented by Dacrydium, Prumnopitys, Dacryacrpus and Podocarpus) (Fig 7). These PCA patterns concur with observations of the pollen diagrams, many of which show a consistent inverse relationship between tall canopy trees and raupō (Fig 6). This relationship is not unexpected as deforestation at lake margins would have created both habitat space and a suitable light regime for raupō communities to expand (Fig 8).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 8. Schematic representation of raupō response to environmental changes during the last millennium in New Zealand (upper panels) informed by palynology represented as generalised pollen diagram (lower panels).

In the upper panels arrows represent flux of nutrients (pink), sediments (vlu) and contaimnants (black). Pre-EMS, EMS and EE phases as described in text.

https://doi.org/10.1371/journal.pwat.0000240.g008

Other processes may also have been involved, separately or in combination, depending on individual site characteristics and disturbance scenarios. From observations of the ecology and morphology of raupō (Section 2.1) and with modern observations of the genus Typha throughout the world (Section 2.2.2), we can speculate that enhanced nutrient and sediment fluxes accompanied catchment deforestation and that these processes also helped promote the spread of raupō in lakes across Aotearoa New Zealand (Fig 8). The wider Lakes380 dataset, when complete, will enable further analytical work to test this hypothesis

Regardless of the exact mechanisms, it is apparent that in dynamic geophysical settings, which are common in Aotearoa New Zealand, these disturbance processes can occur frequently though irregularly in the natural environment, but they have been accentuated and become more pervasive once human activity begins to take effect. So much so that in essence raupō, although a native indigenous species, has become a human associate species in Aotearoa New Zealand and may be useful as an anthropological indicator in palaeoecology (Fig 5A, 5B). Raupō consistently exhibits a close, covarying relationship with Pteridium, the single most ubiquitous paleoecological indicator of human activity in Aotearoa New Zealand prehistory, primarily due to its favourable response to forest clearance by fire [110]. As with Pteridium, raupō is not a perfect indicator of human activity as reflected in the category 1 sites where it rises to prominence before independent evidence for human activity. In most of these cases, however, raupō prominence is short-lived and likely to be a response to episodic disturbance in contrast to the sustained expansion phases seen in the human era.

At many sites, the raupō response to these disturbances is tightly coupled with that of Pteridium, while at others a lag between the two is evident (Figs 4 and 5), pointing perhaps to local environmental factors and differing types or scales of human activity between the sites. These different raupō responses may provide clues as to the exact processes or antecedent site conditions occurring at the time and may also serve to inform wetland management plans that involve raupō (see 5.3). For example, Lake Kereta is a shallow and narrow basin, formed in a small interdune hollow that parallels the western Northland coastline. This setting and configuration is clearly favourable to raupō expansion following disturbance during and prior to human settlement (Fig 6) and could be usefully factored into future management plans for the site.

5.2. Spatio-temporal patterns of raupō

This new dataset reveals some interesting geographical patterns in the timing of raupō expansion during the past ~1000 years (Fig 9). Most of the sites where a raupō expansion does not occur are located at higher elevations, presumably close to or beyond the natural limits to its current distributional range (Fig 9A). A few of these sites are also coastal localities however, suggesting that tidal influence may have promoted competing salt-tolerant species. Nevertheless, some localities that are proximal to the coast also dominate the 13 category I sites, where raupō expansion is observed before the rise of typical paleoecological indicators of early Māori influence (Fig 9B). While localised natural disturbance events such as storms may be a likely trigger of these early raupō expansions, the intriguing possibility emerges that they could also be atypical examples of early Māori settlement impact, given the strong overall correspondence between raupō and human disturbance.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 9. Spatio-temporal patterns of raupō from 92 pollen records from lakes in Aotearoa New Zealand during the past ~1000 years.

Land elevation above 800 m a.s.l shown in grey. (a) sites where raupō pollen is absent or in trace amounts; (b) category I where raupō expansion is observed pre-EMS; (c) category II where raupō expansion is first observed concomitant with or during Evidence for Māori Settlement (EMS); (d) where raupō expansion is first observed during the European era (EE). Basemap showing Aotearoa New Zealand sourced from the LINZ Data Service and licensed for reuse under the CC BY 4.0 licence. (https://www.linz.govt.nz/products-services/data/licensing-and-using-data/attributing-linz-data).

https://doi.org/10.1371/journal.pwat.0000240.g009

This human disturbance connection is clearly seen in the category II site distribution which encapsulates the majority of sites (Fig 9C). Sites in this category are distributed throughout the two main islands at lower elevations and also extend to Chatham Islands. This pattern supports the notion that raupō was promoted indirectly and possibly in some cases, directly, by the rapid spread of early Māori settlers throughout most of Aotearora New Zealand. Our new paleoecological dataset shows that the Māori settlement phase was a time of maximum expansion for raupō across Aotearora New Zealand during the last millennia, and by implication perhaps more generally. Although not apparent from Fig 6, most of these records show a subsequent raupō decline during the European era. Multiple factors for this may be at play, but the most obvious cause is the widespread draining of wetlands and associated management of lakes to promote intensified European agriculture. Consistent with this observation is the comparatively small number of sites where rāupo expansion is only observed in the European era (Fig 9D).

5.3. Raupō as a resource in prehistory

As discussed earlier, historical and archaeological records from many parts of the world along with modern observations of traditional indigenous cultural practice attest to a rich association between people and Typha as a cultural resource. These reports and observations include examples showing the significance of raupō to Aotearoa New Zealand Māori, who had clearly developed a variety of important uses for the plant by the time European settlers arrived (Table 1). Today raupō is recognised as a taonga species for Māori [116], meaning that it’s treasured not only for its utility resource value but also in terms of spiritual connection. Our lake records depicting the history of raupō over the past millennium (Fig 9) bring an interesting perspective to these observations. Raupō expansion during the past millennium, commonly promoted by anthropogenic forest reduction or clearances and associated sediment and nutrient flux, would in turn have provided new opportunities for human exploitation of this valuable food and material resource, underlining further the human associate argument.

Combining this resource exploitation, along with the raupō pollen records that depict close coupling with human activity, raises a further question: was raupō expansion actively and deliberately practised by prehistoric Māori or was the association based on serendipity and opportunity, or both? Evidence for the former certainly exists, such as the deliberate introduction of raupō to the Chatham Islands in the early 19^th century [6]. Our records also show several sites in southern, upland regions that approach or perhaps are even beyond the natural limits of raupō, yet raupō pollen curves show expansion during the early EMS phase followed by decline at sites like Lake Chalice (757 m a.s.l), Horseshoe Lake (681 m) and Lake Heron (692 m.; Fig 5B).

5.4. Raupō: Friend or foe?

It could be argued from observations of modern disturbed wetland environments, supported by the results we present here, that raupō is an invasive species in Aotearoa New Zealand, consistent with suggestions about recent Typha expansion in North America [2]. Whilst its capacity to rapidly colonise and spread across disturbed wetlands and lake margins promoted by increased sediment and nutrient flux is beyond dispute [20], the same attributes can be beneficial in certain circumstances. By intercepting sediment and nutrient flux at lake margins, raupō may serve as a buffer to potential contamination of a lake ecosystem. Preliminary results from Lakes380 research suggest that this may have been the case at least at some sites during the early EMS era. For example, at Lake Horseshoe, Canterbury, the marked increase in raupō during the EMS era is accompanied by distinctive changes in the diatom flora including a decline in pelagic taxa (particularly Discostella stelligera) and an increase in benthic/epiphytic/tychoplanktic taxa (Pseudostaurosira brevistriata and Staurosira construens) [117]. These algal changes are consistent with lake shallowing but also with increased water clarity estimated from diatom assemblages. Subsequently, during the European era, these changes are reversed and accompanied by other major compositional changes in the diatom flora that indicate progressive deterioration of water clarity. At the same time, raupō pollen levels decline progressively towards the present. Whether this recent raupō decline is a response to excessive deterioration of the lake ecosystem or due to coincidental landuse practice cannot be determined from these results and both scenarios might apply. Nevertheless, these changes are consistent with modern ecological observations of raupō-rich littoral zones acting as both a sediment trap and nutrient buffer, with variable effect dependent upon landuse activity in the catchment. Our palaeoecological data suggest raupō is an “autogenic ecosystem engineer” [118] with capacity for transitioning a lake system towards the oligotrophic end of the nutrient spectrum.

These various observations from palaeoecology bring an additional perspective to the idea that raupō can play a role in bioremediation, whilst archaeological and anthropological records attest to its wide-ranging value as a natural resource. To ignore its capacity for providing these benefits due to its invasive threat is to overlook a potentially valuable range of ecosystem services. Rather, there needs to be a robust assessment of a spectrum of ecological scenarios under which raupō may or may not be able to perform a remediation role in managed wetland recovery.

A recent example of this approach is a study of Lake Oporoa, one of the 92 sites included in the current investigation. At this shallow lake in southern North Island/ Te Ika-a-Māui with strong cultural heritage [119], paleolimnology and mātauranga-a-iwi (Māori indigenous knowledge) have been applied in tandem to trace historic and prehistoric trends in water quality and lake ecosystem health. At Lake Oporoa, raupō rises to prominence during the EMS phase (i.e., category 2, cf. Fig 6 accompanied by notable ecological changes in the lake, recorded in both the diatom and bacterial communities. Nevertheless, a more pronounced ecological shift in Lake Oporoa coincided with rapid catchment deforestation and conversion to pasture following European settlement. Further deterioration in water quality has occurred since approximately 1960 CE, when synchronous increases across all proxies indicated enhanced productivity and periods of anoxia. Raupō declines during this latter EE period after prominence in the EMS and early EE. Local Māori place high value on both its cultural significance and its role in lake ecosystem health (https://lakes380.com/lake_story/whakahokia-te-mauri-o-oporoa/), and as a consequence raupō features prominently in a proposed phytoremediation plan to restore lake water quality informed by paleolimnology and mātauranga [119].

The distinction between raupō as “friend or foe” is in large part a value judgement and, whilst beyond the scope of this paper, is now the focus of a more extensive investigation by the Lakes380 team that aims to determine the impacts of raupō expansion phases in the past using various water quality and trophic level proxies. This rich archive encompasses a range of catchment, hydrological, and lake ecosystem settings, enabling the development of scenarios under which raupō may provide net beneficial ecosystem services.