Final Report

Final report on 2019/20 Trial of Site-based non-chemical weed control for ecological restoration

Many findings have already been reported, some in more detail, in earlier photo observations and/or text reports. Here we try to summarise our experience of the application of this methodology to a site during a single year of integrated ongoing survey, planning and intervention.

Some locations are named, eg Zone Ca, or ZoneCaKSS. Links to all the Trial Zones and locations, or “sub-zones”, are provided here.

This trial aimed to

  1. demonstrate the effectiveness of manual weed control over time
  2. help establish optimal rates of weed reduction for gradual replacement of invasive plants with native wild vegetation.

    NB Maximum effectiveness means maximum growth, density and diversity of native vegetation. Weeds are removed because they are inhibiting existing or potential native habitat. So effective weed control requires assessment of the likely or potential development of native habitat, the needs of existing or potential species in that habitat, and the minimisation of soil moisture loss, habitat disturbance, erosion, soil contamination and water pollution.

  3. Assess changes since the fulltime volunteer restoration project of 1997-99, especially observable results of both intervention and non-intervention

Unless otherwise noted Observations below are limited to the defined Trial area, ie the banks below Kaipatiki Roadside, down to the stream and up the opposite bank to the forest path (“Native Plant Trail”).

Ongoing survey

The methodology’s attention to species identification throughout the trial helped preserve, at least temporarily, some existing native plants and invertebrates, and facilitated the beginnings of regeneration in some areas impacted both historically and recently by Reserve users, along the roadside, on the banks below, and along the forest path.

1:1 Plant identification

Of special interest among species identified during the Trial were some natives often either damaged accidentally, treated as weeds due to their common occurrence in cultivated land, or misidentified as exotics they closely resemble.

Following standard practice, no plant was damaged unless it had been identified as exotic. Photographs were taken, sometimes many times until diagnostic features were successfully captured in the images uploaded to iNaturalist for discussion and species ID or confirmation by experts, particularly of ferns. Visitors made some identifications of large trees. No plant material needed to be collected for ID purposes, though some known exotics, eg a fan palm seedling and an invasive grass, were collected for closer study and photography.

With this invaluable help, we learned to recognise some species new to us, eg:

  • the invasive fern Cretan brake (Pteris cretica), which was removed from several streambank locations, releasing native Blechna and Deparia
  • the native Dark nightshade (Solanum opacum), found both beside the path and on sun-exposed edges of the streambank, sometimes among groups of the very similar Black nightshade (S. nigrum), which is itself a benign exotic that provides ground cover, shade that nurses native seedlings, and fruit observed being eaten by birds
  • several invasive palms, after discovering the banks of the Kaipatiki Stream now hold several mature Bangalow and a Chusan, numerous Phoenix and Bangalow seedlings, and an unidentified species of Fan palm seedling.
  • Calystegia sepium, x silvatica, a hybrid of the native Hedge bindweed (Calystegia sepium subsp roseata) with the invasive exotic Calystegia silvatica, ie the locally common weed often referred as convulvulus. This bindweed overtook about 20m of the 1998 roadside planting of kanuka and is still common here, though less abundant in the now-canopied areas of roadside, and possibly not harming the present vegetation.
  • Off-site, we discovered the uncommon native bindweed itself, (C. sepium subsp roseata), possibly planted, in Eskdale Reserve (Domain Rd raingarden), and native Coastal Morning Glory (Ipomoea cairica) wild in Eskdale Reserve, but did not find either species in the Kaipatiki Creek Trial area. [Both C. sepium subsp roseata and Ipomoea cairica appear to be suitable replacements for vine weeds in their natural role of shading and sheltering the outer margins of coastal forests. Establishment of these vigorous but locally uncommon native vines might reduce invasions not only by sun, and wind, but by weeds, refuse and people].

1:2 Habitat

Native fauna observed on the site included:

  • White-faced heron, tui, N. Island fantail, Silvereye, Grey warbler, Kereru, NZ Kingfisher, White-cheeked Rosella
  • NZ Glow Worm, Giant earthworm (Megascolecidae), Cave weta, Tanguru chafer, Nurseryweb spider, Green Orbweb spider, NZ Giant centipede

In contrast to 1997-99, no copper skinks were observed. One fleeting observation of movement on the ground suggested, by its size, a plague skink.

A variety of native and exotic invertebrates were observed throughout the year in all areas of the site, especially in rotting wood (standing or fallen), among dead leaves and beneath undisturbed Tradescantia. Earthworms became common above ground in winter in decomposing Tradescantia piles, and Giant earthworms (native?) were seen on the soil surface when turning Tradescantia piles and underneath discarded plastic weed bags.

We did not see any Tradescantia beetles, or the leaf damage typical of the Leaf-eating Tradescantia beetle. However, no systematic searches were made.

Wide areas of impermeable inorganic materials left in the forest, while hindering successful development of seedlings and juvenile trees, have become invertebrate habitat. During the staged gradual uplifting, and later winter removal, of large plastic weed bags from the roadside and lower streambanks, we unexpectedly exposed a young Cave weta, some centipedes, and a giant earthworm, Cover was immediately replaced with plant material, which may or may not have provided equivalent humidity and predator protection.

The removal of plastic ground cover during unrecognised extreme soil moisture deficit, followed shortly after by a return to drought, may thus have resulted in loss of invertebrate individuals or populations.

Elaeagnus berries lost in the reduction of Elaeagnus invasions would have provided much-needed bird food during the drought.

Pest and predator animals

Domestic cats were occasionally observed in the forest. Feral cats were not observed, but are present in neighbouring forests and are seldom obvious.

Reduction of Elaeagnus would have resulted in partial loss of the cat-proof habitat provided by Elaeagnus to native fauna.

Rat traps set by a volunteer were informally monitored in two locations (ZoneCaWSS, ZoneBdWRS), and several successful kills noted.

Bait and monitoring stations set by contractors were noted and avoided.
Vespula and paper wasps were rarely seen after March 2019.

1:3 Water quality and Freshwater ecology

Observations of the stream were not systematic or prolonged, and an evening visit was not made, but nonetheless we were surprised at the ongoing absence of bubbles, water splashes, or shadowy movements suggesting fish and eels.
Only one or two damselflies were observed, no wasps or bees drinking from the edges of the stream.

This may be partly attributable to water temperature, which was not tested. A single heavy rain event was observed during the winter, in the Trial area, and the flood level in that event was modest, causing little disturbance of streambanks, compared with levels seen regularly in 1997-1999.

During the heavy rain event of 1 April 2019 we observed what the apparently routine (local residents report that the distinctive smell occurs regularly after rain and can be smelled from surrounding streets) overflow of raw sewage into the stream from the many sewage manholes along the streambank. We were unable to obtain information from Council or Watercare on the water quality test results that are done from a weir a few hundred metres downstream of the Trial site.

Untreated stormwater is still piped directly into the stream from private properties and roading. As in 1997, this results in the stream becoming whitish-grey after first-flush rainfalls, which wash heavy metals and other pollutants directly into the stream.

We speculate that following extreme drought, the low level of water in the stream will result in even greater concentration of the usual post-rain, or “first flush”, stormwater pollution. Such concentration of sewage and stormwater pollution in post-drought rain events may already occurred prior to March 2019.

Outside of heavy rain events, stream water was clear in all seasons in the areas accessed during the Trial, but not closely observed further up or downstream inaccessible due to either dense native revegetation, eroded banks and fallen trees, or dense infestations of crack willow and Japanese honeysuckle.

“Willow Crossing” area was, in 1997-99 a slow-flowing area, unshaded on the roadside bank and prone to algal blooms in summer. This area was not observed during 2019-20, as the 20 year old planted trees and new vine invasions now hide the stream and prevent easy access from the road. However, this suggests the dense vegetation is shading and cooling the stream as intended, and algal blooms may have diminished.

Shortly downstream of Willow Crossing, the streambed has, as in 1997, several metres of concrete streambed including a weir.
By chance we observed a regular scheduled visit for water quality data collection, but were unable to contact anyone with information about the results, which we are told include regular automatic monitoring of flow rates.

Stream formation
Sediment in the Trial area’s streambed did not seem (in wading) significantly deeper than in 1997-00, as we recall it. Banks appeared somewhat higher and in places unstable, with more exposed tree roots than in 1997-00, suggesting scouring. There are still undercuts/overhangs for kokopu habitat, but fewer submerged logs and sticks in the streambed, which (excluding periods of drought) probably carries a larger volume of both runoff and piped stormwater than 20 years ago, and higher volumes of sewage/stormwater overflow from manholes after rain events.

As usual and desirable in such habitat, partial dams have formed from logs, water-borne debris (organic and inorganic) and sediment. The pools observed above such dams were no deeper than 30-40cmD during the 2019/20 Project year, one of prolonged extreme drought.

Large fallen and uprooted trees, eg mahoe and treeferns, hang over parts of the stream, often creating overhangs in the streambank. Such treefalls, in otherwise-intact streamside forest, are part of the natural process of stream formation. The resulting dams and eroding banks create wider, slow-flowing areas and sediment sandstone islets, (eventually covered in ferns, carex, and juvenile trees) contributing in the long term to the meandering of the stream, which in general still follows its 1997-00 path. The overhangs provide habitat for kokopu, which inhabited this stream in 1997-99.

Refuse and litter in the streambed
Other than solid waste, toilet pater and plastic refuse left by the sewage overflow of April 1st, refuse in the streambed was minimal. Most of it appeared to be long-standing, including a foam mattress, car tyre, plastic containers, and a semi-buried section of synthetic carpet.

In contrast with the truckloads of refuse and several car bodies removed from the streambed and near banks in 1997-99, we observed only one or two items of larger flood-borne refuse such as telegraph poles, water tanks, construction materials, furniture and vehicle parts.

Litter, especially plastic drink bottles, was also much less common in the streambed. Though there continues to be roadside litter and frequent illegal dumping, including household lots, butchered animals and bulk containers of hazardous and flammable chemicals, these are largely retained near the roadside by the density of the planted trees.

The stormwater drains in the road may now have smaller openings, filtering larger pieces of debris before stormwater enters the stream. We suggested this in 1999, and were told at the time that that would block the drains with rubbish, and that the collection of rubbish from stormwater drains would be an “upgrade in service” prohibited by cost. We don’t know if this service has since been implemented, or if there are just fewer plastic drink bottles in the gutter and stream because of changed habits among motorists and pedestrians.

1:4 Inorganic refuse in roadside and forest

Following the 4x20 litres of hazardous flammable chemical fluid, found, with a pair of industrial gumboots, in June 2018, and the removal of a stack of 8-10 car tyres in March 2019, dumped refuse and litter at the roadside was ongoing until we completed a pigtail cordon along the roadside. Dumpings reported to council for pick-up included furniture, appliances, timber, plastic sheeting, road signage (probably discarded by vandals when they tired of it) and a household lot of mattresses.

Waterborne refuse
A substantial collection of litter, mostly drink containers, was found trapped by Tradescantia in the rough concrete-and-rock culvert channeling stormwater to the stream from the roadside and playing fields across the road (Zone EaKRS/EbKRS).

Equipment remaining from prior activities
During Tradescantia removal in the forest, old broken plastic pest-control tunnels were found in two locations among dense understorey and with tree roots growing through them. The tunnels were able to be detached and removed, and the nikau roots placed back in contact with moist soil and covered with leaf litter.

Black plastic kleensaks and mesh weed bags found under Tradescantia on lower level banks and streamsides were left in place until winter, to avoid breaking rootlets and desiccating habitat. However, with unanticipated and increasing soil moisture deficit and drought, their removal undoubtedly contributed to the later drought-stress and plant loss.

1:5 Plant health

Tree deaths
We noted in our initial survey numerous dead and declining trees within the Trial site. Most were well below the typical height and girth of adults elsewhere on the site.

Among the dead trees were

  • a majority of the observed tanekaha (see earlier reports)
  • numerous mahoe at the pathside or within 2 metres
  • several kanuka of only 20cm D or less, in locations of at least partial sunlight (and some larger, whose branches and foliage, in canopy breaks, appeared scant but could not be seen close enough to confirm life or death)
  • a ti kouka 10-15mH below the roadside planted trees (Zone CaKRS).

Drought stress
In the roadside area, shortly after beginning exploration and intervention in mid-March 2019, we observed leaf wilt in first one or two, then in increasing numbers of mahoe and kawakawa. We also noted pale colour, small leaf-size and secondary disease/insect damage in karo.

These signs, presumably drought stress, were relieved in mahoe and kawakawa by light rains and/or mulching with Tradescantia, but returned intermittently with gradually increasing severity.

Drought stress was not apparent at this time under dense canopy, at the streamside or on lower streambanks, but its later appearance is discussed further below.

Juvenile trees were abundant throughout the streamsides, forest path and lower banks below the roadside.

Seedlings, sporelings or juveniles were rare below the roadside planting in Zones Ca to Eb, and on some eroded or slipped banks, in the immediate area of large tree or branch falls, and on sun-exposed banks (ie where tall trees were absent)

Despite the presence on the lower streambanks of two other apparently healthy mature tanekaha, tanekaha seedlings were not observed in Witheford Reserve outside the defined Trial area except as two depauperate specimens c. 6cm H, in dry soil under canopy margin at (surprisingly) the roadside near a large healthy tanekaha.

We found very few seedlings, exotic or native, in the deep Tradescantia throughout most of the Trial site roadside bank. A few kowhai, kawakawa and Coprosma that were found were at the outer edge of the roadside planted tree canopy, within a few cms of the regularly-mown grassed roadside verge (where grass has now been replaced by a few dried herbs in dust). Just under the dripline bordering mown grass, soil moisture and humus were greater than in the soil further under the canopy, perhaps due to the build up of humus in mown grass.

In late winter and early spring when the ground surface became wet, Tradescantia was uplifted from small patches (eg 0.25 - 1 sq m) under roadside mahoe, kanuka, and karo. This disclosed dozens of newly germinated seedlings of scattered coprosma, abundant kawakawa, and a few kowhai, but these little patches quickly dried out on the next sunny day, even between days of rain, so the bared patches were covered with loose Tradescantia to prevent further desiccation.

After the year of observation and reflection on all possible causes, we suspect chronic desiccation as a major contributor to both the ongoing tree deaths and the absence of seedlings on the roadside bank, but have been unable to obtain specialist advice or assessment to support or negate that. A limited seedbank and soil flora in this constructed roadside may be another factor.

Native vines
Several native vines were found entangled with fallen trees and weed plants alongside and overhanging the path, vulnerable to destruction by the public, both in maintenance of path access and due to the common misperception of all vines as a threat to trees. Some of these native vines alongside the path have attained nearby canopy, and even crossed the path where canopy supports it. Our observations of their presence in canopy and the spread of seedlings indicate these vines, mostly generated since 1999, are providing not only bird food but also much-needed shade and weed-prevention.

A Bush lawyer (Rubus cissoides) trunk about 6cm in diameter lies horizontal near ground-level alongside the path, its upright branches and foliage at times overhanging the path and thus requiring regular pruning for the safety of Reserve users. We discovered this vine during our ad hoc clearance of fallen trees, Elaeagnus and moth plant overhanging the path at that point. We were eventually able to remove the weeds and push several Rubus branches back from the path, leaving several young upright shoots of bush lawyer clear of the path and able to climb the steep pathside bank into the canopy. However, without ongoing manual weed-control, ongoing intervention to direct the Bush lawyer’s growth away from the path, and public recognition of the thorny vine and its horizontal trunk, it is liable to be damaged or killed by a Reserve-user, potentially resulting in permanent loss of its fruit and canopy.

Parsonsia and karaeo are present but not common. Young vines climb both weed and native juvenile pathside trees (Elaeagnus, mahoe and Coprosma), with mature vines observed with difficulty in the canopy. Where these vines are within a metre of the path, identification and directing them away from the path will be necessary to avoid their destruction.

Weed control

2:1 Conservation of soil, flora and water

In the dry conditions it was especially important to reduce weed mass only sufficiently to release existing or imminent native vegetation. This helped maintain shade and moisture which would have been lost in complete weed removal.

Hand-control of Tradescantia at the pathside avoided further loss of native seedlings and soil health. No other weeds were seen on the path, but native Carex, Solanum americanum (native nightshade), native Basket grass (Oplismenus hirtellus imbecilis) and possibly fern and treefern sporelings, may have been perceived, as some previously-healthy examples of these were found decaying in initial survey.

A new occurrence of nightshade, possibly native, was found uprooted in March 2020, in a humus-rich section of pathside area where Tradescantia had been removed and composted during the Trial. This nightshade, whether native or exotic, presented no threat to habitat or path users, and would have provided excellent cover for native seedling development. It was near a sewage manhole that overflows after rain events, where vegetation of any kind is particularly valuable in filtering run-off and restraining the distribution of raw sewage and associated inorganic refuse.

2:2 Conservation of freshwater ecology

The catchment of the Eastern arm of the Kaipatiki Creek above the Trial area is steep and largely paved and piped. In 1997-99 heavy rain events produced powerful currents sweeping debris across the footbridge and into trees up to at least 2m above the normal water level.

Eutrophication of the stream is likely if green leafy material enters the stream, which can happen if plants are cut or uprooted and left to decompose in areas of run off or on banks reachable by floodwaters.
Location of piles and any loose leafy material during the Trial was done, then and now, with that in mind. All loosened nitrogenous material was placed above flood levels, in situations where nitrogen-enriched run-off would have been filtered by carbonaceous ground litter such as logs, sticks and dry dead leaves.

Stream levels were monitored after heavy rains, and colour and opacity observed throughout the Project. The above protocols appeared to have been successfully observed.

2:3 Habitat conservation in tree/shrub weed control

Trial interventions considered Elaeagnus and Tree privet seed production, as both seedlings are moderately shade tolerant and quickly become firmly rooted. We decided to leave one large heavily fruiting branch of the largest Elaeagnus tree, to maintain some of the bird food supply.

Most of the Tree privet we treated were juveniles, possibly approaching fruiting, canopied by Tree privet to 10mH, and surrounded by mature fruiting Coprosma.

During the subsequent severe drought, during which many forest fruits have failed to mature, we regretted having removed any fruit-producing material at all. We also observed that in the extreme dry conditions during this trial and to date, leaving more branches and foliage on both tree privets and Elaeagnus would have adequately suppressed the weed trees’ growth while maintaining the shade they had provided for released natives and the forest floor beyond.

2:4 Trial of tree/shrub weed control techniques

The Trial enabled ongoing monitoring of the results of Trial techniques of Tree/shrub weed control by “partial breakdown”.
We found releasing adjacent native plants and reducing weed mass and reproduction by the trial techniques to be less time-consuming and more effective than anticipated.

Tree privet
While more resilient than their native neighbours, these were similarly sensitive to the extreme drought. Consequently their regrowth during the Trial was slower than observed elsewhere in previous seasons, and the very satisfactory results cannot necessarily be expected in wetter areas and conditions.

Nonetheless, the variety of strategies and techniques used for release of natives and gradual attrition of tree/shrub weeds were found to be practical and effective, allowing adjacent native vegetation to thrive and surpass the weed trees until further intervention. (All these tree weeds will of course recover without further intervention).

Tree privet from 30cmH to 4mH were reduced by partially-breaking and bending down branches or in some cases trunks. Follow-up interventions addressed any branches interfering with native regeneration nearby, while maintaining any foliage supporting adjacent vegetation by providing shade and/or ground cover.

Repeat interventions were ad hoc as each specimen was encountered during general monitoring and restoration activities in subsequent weeks or months, with small reductions requiring little effort or time: eg bending down or snapping a branch, or bending all its branchlets backwards, and making larger breaks and suppressions as the tree weakened. As expected, only a few very small ones became sufficiently weak to be easily uprooted during the year, but many large branches ceased to produce leaves or died. Some partially-broken branches or trunks died and were easily broken off. Where this occurred in a large branch or branches, regrowth from other branches slowed significantly.

Specimens in sunlight showed rapid vigorous regrowth, with substantial foliage developing on partially broken branches.
Those in moderately dense shade were slower to regrow, and had been more spindly with less foliage to begin with. Some of these produced very few leaves after their first follow-up intervention.

After several repeat interventions, and during the ongoing drought, all the tree privet juveniles in the Trial area are, in March 2020, considerably shorter (despite all new growth pointing upward), with scant slow-growing foliage, insufficient to impede adjacent vegetation and not expected to need intervention, or to reach a height prohibiting intervention by volunteers, in the next month or so. The continuing regrowth drains the roots’ resources, as the live but unproductive upper trunk and branches demand ongoing hydration from the roots, while supplying no nutrition in return.

[ iNaturalist member baldeagle has explained the plant processes involved, and established that “girdling” kills privets more quickly than felling. Regarding public health and ringbarking, he observed that dead tree privet many metres high stand intact for years, even supporting other falling trees in a storm.]

Elaeagnus reflexa
A tree several metres high with several trunks to 10cmD was partially felled and left hanging by its upper branches caught in the surrounding native canopy. Regrowth shooting from the stump was in shade, and not obvious by Mar 2020, but will continue, assisted by any canopy loss among the adjacent native trees.

Care and gradual staging were used to avoid the weight of large cut Elaeagnus branches falling on native vines hanging among them. Cutting or ringbarking side-branches of such Elaeagnus and waiting for them to decay gradually reduced their weight so that the eventual removal did not damage the native vines.

Many smaller trees, some supported by native trees or vines, were gradually reduced by the same method as described for Tree privet. Seedlings unable to be uprooted were broken off at base, or partially broken and bent down. 3mH Elaeagnus entangled in native vines were treated by partial breakdown aimed at releasing adjacent habitat and slowing their growth sufficiently to ensure the native vegetation was unhindered.

Several Elaeagnus to 2-3m H, or sprawling, were observed in August 2018 (pre-Trial) under the big pine tree (ZoneBdKRS) were reduced then and found without significant regrowth in March 2019. Many more throughout the site were addressed as found from March-June 2019. Slender specimens to about 2mH showed little or no regrowth by September 2019 and received no further intervention. Larger specimens, eg 3-4cmD, had several follow-up treatments, all simple and brief, leaving small plants reduced to a single stem with any small flexible branches gathered and bent into a loop, and large plants leafing only on one or two stems, which were easily partially broken.

Our experiences with the technique elsewhere (Eskdale Forest paths 2009-12, and Gahnia Grove 2018-29) in less-intensive projects we would plan monitoring and intervention quarterly to annually depending on the conditions and the situation and vigour of the specimen, suggest follow-up monitoring and intervention are initially required quarterly to annually, depending on the micro-climate and the current vigour of each specimen.

Tree/shrub weed seedlings
Few young tree/shrub weed seedlings were seen until summer, and then with a high attrition rate despite the removal of competing Tradescantia. Their attrition was likely increased by the exceptional drought, as shade by this time was not deep due to cutting (by reserve users) of pathside vegetation and leaf loss by drought-stressed trees. The surviving weed seedlings were almost all Tree privet. Judging by their number and survival rate along the lightly forested streambanks, where Tradescantia is absent or sparse we expect thousands more to arise and to develop once soil moisture recovers.

Due to their simultaneous emergence in groups of hundreds and thousands towards the end of the Project, these weed seedlings were not counted, but some particularly striking occurrences were recorded in photo observations.

2:5 Exotic Palms
It was a shock during initial survey (March to August 2019) to find several mature invasions of Chusan and Bangalow palm on the streambank. Numerous Phoenix palm seedlings to 50cmH were easily suppressed by looping and knotting. After some months, the first specimen thus treated died, leaves green and intact, its base and root entirely rotted.

Specimens to 60cmH were observed elsewhere in the wider site during survey, suppressed in the same way during survey, have not been monitored but could be easily located for assessment if required.

Dozens of smaller exotic palm seedlings observed were uprooted, both on forested banks and in the streambed itself, and are expected to be increasingly numerous in this and subsequent autumns, with several bangalow now established along the streambank probably maturing and fruiting this year. We expect the current dry conditions to favour exotic palms over nikau and other native trees.

2:6 Vine weeds
Occasional Moth plant seedlings were found and uprooted on the streambanks and in or beside the path. There are about a dozen pod-bearing infestations at the roadside, and a large streamside infestation throughout the canopy of a mahoe 10mH, just downstream of the Trial area.

A wide invasion of honeysuckle and a few Moth plant vines hung from several kohuhu, karo and mahoe along the roadside at the top of the mostly-treeless streambank below.

Mature moth plant vines look very similar to mature Parsonsia heterophylla vines, and the two are sometimes tangled together in the same tree, so great care had to be taken to ensure foliage was identified before cutting a Moth plant stem.

Beside the forest path, Moth plant hung entangled with Parsonsia or Rubus cissoides. (Here, the Rubus cissoides vine was also entangled with Elaeagnus, whose rough brown stems superficially appeared similar).

Intervention in these cases was by gradual reduction over several months, as parts of the invasion could be located, identified and removed without damage to native vines.

Post-trial, regrowth is expected in any remaining undiscovered parts of such vines.

2:7 Tradescantia

Initial survey of Tradescantia in the Trial site started, after a week of considerable rain, in March 2019. The first area surveyed in depth was the roadside bank in Zones Ca to Da, where a small exploratory removal had been done in Oct 2018, and some photos included our RENH funding application.

Here in 1997 both the roadside and the “lower level” held little but a few juvenile ti kouka and dense kikuyu.
The kikuyu was mown at the roadside and uncontrolled beyond. The roadside kikuyu was sprayed prior to the planting, and all kikuyu throughout both roadside and lower level was eventually deeply mulched by volunteers with woodchips. In March 1998 the roadside was densely planted by about 100 community volunteers, with ecosourced native tree seedlings purchased by North Shore City council. In Zones Ca-Da, the planting extended about 6m from the roadside, sloping downhill.

The lower level behind was never successfully planted.

Hundreds of metres of woodchips were donated directly to the volunteer restoration project, with Council approval, by a local treecare company. This was the first use by North Shore City Council of wood chips as mulch, as woodchips had been considered to transmit weed invasion. Technical advisors and observers were unanimous in their commendation of the results of the tree planting, and no new or increased weed invasion was observed.

NSCC had until then been importing commercial bark mulch from the Waikato, at great cost. The bark mulch, coarse and smooth, often failed to aggregate or decompose before being rapidly lost from its intended site.

In 2019, as in 1999, Kaipatiki Rd was wide and busy, guttered and drained, and bordered by an informal footpath about a metre wide, of recently-mown mixed exotic grasses.

Roadside Trees
The roadside margin of the planted canopy was dense enough in places to discourage the entry of passing pedestrians, the foliage of the closely planted trees occlusive from ground to canopy in some places. Gaps in this roadside canopy margin sheltered a few juvenile natives (mapou, ti kouka, Coprosma, mahoe, mamangi), probably wild, and several small invasive trees and vines (Tree privet, woolly nightshade, Elaeagnus, honeysuckle, moth plant).

There was dense, deep ground cover by Tradescantia throughout the roadside planting in Zones Ca-Da, ending at Zone BdKRd/ KRL, near the entrance to the well-used walkway connecting Glenfield College and Kaipatiki Rd to the Witheford Dr/Bayview neighbourhood.
[In ZoneBdKRd the plantable space was, and remains, limited to a steep bank 1-3m wide, above 20m of 2mW path entry almost parallel to the road. This part of the 1998 native planting, bounded by foot traffic both above and below, holds diverse trees and shrubs and ferns, perhaps all now wild, the taller trees often leaning over the path so being pruned or felled, and those nearest the path entry appearing stunted, presumably by soil compaction, lack of canopy and ground cover. Several juvenile wattle, woolly nightshade, Elaeagnus and blackberry were controlled among these trees. The roadside lower branches of Coprosma, mapou and mahoe obstruct recreational walkers and access to cars parking daily opposite the College, so are prone to damage. Pruning back to about 1.2m from the road successfully avoided further damage to the trees and increased the density of roadside foliage].

Lower level
About 6m below the roadside, the sloping bank ended in a flat “Lower level”, several metres wide and ending at the top of a steep bank. This level was similarly covered in dense Tradescantia.

There were no trees or understorey on this lower level below the roadside planting, except a single vigorous but juvenile nikau, and a large old multi-trunked ti kouka, deep in ti kouka leaf litter. This stand was about 10m-15m H with sharply leaning and bent trunks, of which one had recently fallen, one piece several metres long remaining for months high in a group of trees at the nearby path edge.

Steep bank
A steep clay bank below the “Lower level” held moderately dense wild native trees (rawirinui, mahoe, rewarewa, ti kouka, mapou, ponga, wheki, a single live tanekaha and another now dead) and understorey (kiekie, coprosma, kawakawa, pate, ferns), several woolly nightshade over 5mH and several clumps of Kahili ginger. This steep bank was less densely covered in Tradescantia than the roadside or lower level; bare clay was visible among root buttresses and around a decaying uprooted tree base with fallen trunk and branches up to 2mD (observed already long-decaying in 1997).

Flat Streambank
At the bottom of the steep bank was a flat sedimentary sandstone streambank, canopied directly by several tall close wheki, and more distantly by taller canopy trees on the surrounding banks. The streambank was densely covered in Tradescantia, several Kahili ginger stands, a few scattered juvenile treeferns, deep piles of fallen fronds around the wheki.

A stand of Palmgrass (Setaria palmifolia) was easily broken down or uprooted, and suppressed with wheki fronds. There was some regrowth by February 2020.

The ginger was suppressed by partial breaking and bending down of all stems, covering them deeply with about 1m3 of Tradescantia. This trial technique proved simple and effective. Several months later the Tradescantia was removed, to find the ginger roots had rotted, and several large masses of otherwise-healthy and viable tubers were easily uplifted and removed to the roadside. A smaller amount of tubers remains, and would be likely to respond to a repeat intervention, but would need more interventions and monitoring in the current absence of Tradescantia to suppress stems and assist rotting.

The streambank edge held the same density and depth of Tradescantia, but here the Tradescantia hid young ferns, mostly Pteris tremula and Deparia, which grew vigorously after release.

Sporelings of treeferns are developing on the flat bank beneath wheki. A kiekie forms a wall about 2m H x 6m across one end of the streambank, descends the bank to the water, and climbs several metres up the steep bank and further up trees.
A group of young ferns in its shelter flourished on release form Tradescantia, and native seedlings incl. kanono emerged in Spring. The development of all these, and possibly their survival, has now been restricted by drought. There were signs of recovery among them after light rain in March.

The trees on the roadside and on the steep bank, and three tall trees nearby, (a radiata pine to the NE, and a tanekaha and tree privet on the opposite side of the stream), all contributed to the thin and translucent partial canopy of the lower level, steep bank and streambank. A substantial proportion of this fragile canopy was, and is, provided by a single Tree privet c.10mH, leaning over the stream and apparently outcompeting an adjacent mahoe, the latter of similar height but with only a few short branches and scant foliage.

In March 2019 all the above areas were covered from roadside to streambed in deep dense Tradescantia.

The trial began with:

  • Assessment of canopy (broken in places, mostly thin and somewhat translucent, of mahoe, ti kouka, pate and treeferns; dependent on several woolly nightshade and the large tree privet)
  • exploration of dense deep Tradescantia for seedlings and sporelings (no seedlings; many Lastreopsis on the steep bank above the stream, to 30cmH and completely hidden under the Tradescantia),
  • identification of mini-sites for monitored partial Tradescantia removal (in wet areas around run-off channels around several stands of tall old ti kouka, where Tradescantia was mixed with many layers of ti kouka leaf litter)
  • uprooting of Tradescantia along a strip of ground about 80cm wide x 5m L on the lower level, placing the loose material at the outer edge of roadside canopy to suppress kikuyu, Paspalum, Yorkshire fog and Cynodon dactylon, retaining moisture and supporting revegetation during its decomposition.
  • removal to the roadside of a pile of 8-10 recently dumped car tyres from under one mahoe, and from the wider area two wooden pallets, some iron pots, and a few pieces of timber and plastic refuse, most previously hidden by the Tradescantia under the roadside trees.

Roadside bank - soil moisture deficit and drought stress

A day or so after these activities on the roadside bank, intense sunlight returned, penetrating the light canopy and rapidly drying any spots bared by removal of Tradescantia or refuse. Where the tyres and Tradescantia had been removed, some mahoe leaves drooped (on the canopied side of the tree, not the sun-exposed side; perhaps because surface roots on the exposed side were already adapted to dry soil).

The removed Tradescantia was redistributed to cover the areas exposed by rubbish or Tradescantia removal. After about 10 days and a little more light rain, the affected mahoe leaves appeared normal.

Over the following week additional Tradescantia was brought from the streambed and applied as mulch to the recently bared areas, supplemented by harakeke prunings found nearby.

Among the root buttresses of a mature karo 1-2m from the kerb, loose dust could be penetrated with a finger to a depth of 1cm.

Closer inspection of the soil surface showed tiny white mahoe rootlets in the soil beneath the Tradescantia, exposed anywhere Tradescantia was removed.

We speculated that the soil moisture deficit had induced mahoe roots to emerge or increase at the surface to access moisture from dew and intermittent light rainfall retained by the Tradescantia.

Some observations indicating drought stress among the roadside trees after removal of some refuse and small areas of the previously deep dense Tradescantia.

After these removals and subsequent observation of drought stress, as much Tradescantia as possible was carried up the bank from the streamside to mulch all the roadside trees as deeply as possible.

Thereafter, Tradescantia removal was limited to the following areas:

A. Moderate shade under tall canopy

Under large old pine, kanuka and tanekaha, where leaf litter was longstanding, deep and dense, Tradescantia cover was initially light, superficial, and maybe relatively recent. Tradescantia was removed completely from these areas, with no drought-stress or other impact observed, and released juveniles and seedlings continuing to grow. Few new seedlings or sporelings were observed until March 2020, and are not numerous now, in the undisturbed deep litter.

B. Quadrats of Tradescantia removal

Several Quadrats of Tradescantia removal, each about 80cm sq, were marked with ground staples and plastic tape in the midst of deep dense Tradescantia under the roadside trees, from 1 to 5 metres from the road.

A few dead leaves and pieces of kanuka brush were placed in each to provide micro-shade. These quadrats remained dry and bare throughout the year, with only one or two producing one or two seedlings of Coprosma or kawakawa, and a few Crepis capillaris (benign exotic) seedlings, all of which died soon after.

Several smaller (c.40cmD) similar areas were briefly released in winter, releasing vigorous new seedlings of scattered coprosma, dozens of kawakawa, several toatoa (Haloragis erecta) and a few kowhai and possible mahoe. These newly-bared areas quickly dried out when the sky cleared, so the seedlings were unlikely to survive and the areas were re-covered with loose Tradescantia to prevent further stress to nearby trees.

C. Streamsides

Tradescantia was carefully removed from several areas of streamside with dense lush native vegetation. In these areas few seedlings were found other than those already evident emerging from light Tradescantia coverage. The released juveniles and large seedlings (eg 10cmH) flourished as expected, at least until late summer, when some individuals started to be affected by drought.

Throughout a level sedimentary sandstone bank less than a metre above water level in zone CaKSS, moderate regrowth of Tradescantia had occurred by March 2020s. Strands, gently uplifted, revealed dozens of ti kouka, several coprosma, a few puriri and mahoe, and some carex seedlings, with numerous unidentified fern sporelings.

D. Steep bank above stream

As significant rain resumed in winter, Tradescantia was also gradually removed from the steep bank above the stream, and eventually all carried uphill to reinforce the mulch under the roadside mahoe, kawakawa and karo.

Numerous patches of the creeping native fern Lastreopsis hispida were found under deep undisturbed Tradescantia in March, and left undisturbed to avoid moisture loss. They were gradually released from the Tradescantia with the return of steady rain in winter, and grew rapidly. The Lastreopsis were joined in spring by new sporelings.

Some Lastreopsis grew to c. 1mH. Some of the other sporelings (ponga, Pteris tremula, Deparia) developed to c. 30cmH.
Many were still vigorous in March 2020, but some may not have survived the Spring/Summer drought without Tradescantia cover.

E. Forest pathsides

Where the path had been widened through plant damage and trampling, Tradescantia was absent or reduced, leaving compacted and bared clay holding a reduced number of juvenile trees and the exposed above-ground roots of numerous mature mahoe and pate.

After on-site discussion with Council and contractors, the path was re-defined using ground staples and plastic marking tape at ground level. Piles of Tradescantia were applied outside the taped edge, to restore soil health and moisture and to define the path for Reserve users. It was hoped that the remaining juvenile trees would then be permitted to grow freely with vertical pruning only, with ground covers (carex, Basket grass, fungi, mosses etc) recovering or returning, supporting soil health and hydration for the health and ongoing development of the native trees along the path.
We considered creating signs to educate Reserve users about the need to support shade and canopy development, and the role of shade in preventing weed invasion. However, we found we had not the resources to do this thoroughly and effectively.

The pathside Tradescantia piles were monitored at least weekly through winter, and regrowth scooped back onto the piles, with some turning of small piles to restrain growth and accelerate decomposition, to avoid requests for herbicide spraying.

Piles were gradually combined and reduced through Spring and early summer, preventing regrowth and to permit sporeling/seedling germination in released areas of pathside.

Due to the shortness of winter, Spring/Summer drought and increasing soil moisture deficit, the Tradescantia along the pathsides had mostly decomposed by early summer. Exposed and compacted clay along the pathsides remained bare. In areas of leaf litter, humus or light Tradescantia regrowth, generally around the bases of trees or rotting logs, native seedlings began to appear during the summer (Coprosma, pate, mahoe).

In March 2020, many pathside Carex partly or completely yellow or brown in May 2019 are growing healthy new leaves. Those with Tradescantia piles beside them are vigorous and lush. Basket grass is increasing from remnants in the path edges.

Some pathside juvenile trees have grown slightly, some are little changed. Drought stress is common among the broad-leaved trees, especially mahoe, hangehange and kanono.

Of three rawirinui juveniles observed growing vigorously beside the path in June 2018, only one survived the pruning and felling of pathsides in early 2019. This one, set furthest back from the path, is growing well.

The horizontally-pruned and sun-invaded stretches of pathside have lost some of their smaller trees and ferns to drought; particularly hangehange, pate, kanono, mahoe, and their epiphytes, including petako and huruhuruwhenua.

Many of the juvenile trees and ferns epiphytic on pathside trees have grown increasingly pale and flaccid during the summer, with some unlikely to recover.


3:1 Plant health and reproduction

After the year of observation and reflection on various possible causes, we suspect chronic desiccation as a major contributor to both the tree deaths noted in Initial survey, and the absence of seedlings on the roadside bank.

A limited seedbank and soil flora in this constructed roadside may also be a factor in the absence of seedlings, but note that Coprosma and kawakawa seedlings did germinate both under Tradescantia and in patches released earlier from Tradescantia, but none survived.

The several juvenile kotukutuku (Tree fuchsia) along the forest path became vigorous and leafy during winter, before losing most of their leaves in Spring.

“Adult kotukutuku #1” (ZoneFbWSS), in an area of run-off that remains largely Tradescantia-covered and relatively well-hydrated, produced sparse foliage and flowers, but the foliage blackened and crumbled early in summer, possibly due to the tree’s advanced age, and perhaps poor root instability and ill-health in an area of sewage overflow and stormwater run-off after heavy rain.

3:2 Climate change, soil moisture deficit and low rainfall

Over the course of this trial, we observed and struggled to comprehend plant responses and conditions previously unknown to us.
In late 2019, as the Australian bushfire disaster worsened, news media began to provide background on the extent of climate change and drought in that country. We then learned that the effects of the earlier Antarctic warming and Extreme Postive Indian dipole events contributed not only to the Australian drought and subsequent fires, but to the current ongoing drought in the North and East of New Zealand, and such climate events are expected to occur with increasing frequency with ongoing global warming.
We then understood our observations and concerns to reflect the unprecedented depth of soil moisture deficit in the currently ongoing 2019-20 hydrological and meteorological drought.

“The lowest rainfall for summer was recorded in Dargaville (63mm), Leigh (59mm), Whangaparaoa (60mm), North Shore (53mm), Western Springs (62mm) and Auckland Airport (73mm)”.

We also noted throughout the year that Metservice temperature and rainfall records reflect Whangaparaoa, not Glenfield, where rainfall was often less or completely absent, and maximum daily temperatures on North-facing ridges were informally recorded up to 10 degrees higher than temperatures reported by weather stations.

Locally at least, the climate and soil conditions are very different from even recent years. The loss of even small amounts of groundcover, shade or shelter can cause critical soil desiccation, tree death and habitat loss in these conditions. If ground cover and/or shade are lost and cannot be restored, we envisage escalating desiccation leading to devegetation, erosion and land slips, particularly on ground contoured by road construction using loose porous materials, and without run-off due to road guttering and piping directly into the stream.
We believe the impact of soil moisture deficit and continuing drought on the Trial site was exacerbated in some places by both interventions for amenity, and by our own manual weed removal and the removal of large areas of plastic and other inorganic refuse.
In these conditions the Methodologies moisture-retention techniques, including the use of all material as mulch, failed to completely mitigate drought-stress.
From March to at least June, intense heat and sun recurred between rainy periods, immediately drying all surface moisture in the roadside area, and on lower banks arresting the development of seedlings both native and exotic.

Ongoing drought reduced canopy density through leaf wilt, leaf drop, smaller leaf size, and in some cases tree death.

Piles of Tradescantia brought uphill from the lower streamside banks quickly lost most of their mass, leaves completely disappeared (helped by earthworms and other invertebrates) and the remaining stems did not shoot, instead drying and disintegrating, producing little humus.

Piled Tradescantia not only failed to regrow and spread beyond the pile, but much of the remaining Tradescantia alongside it stopped growing, even in winter.

By January 2020, areas of undisturbed Tradescantia had also become sparse, letting light in and moisture out, and on the drier banks became discoloured (purplish) and wilted.

Our reduction of the relatively small amount of vine and several juvenile tree privet at the roadside of the Tradescantia trial area likely contributed to tree stress by letting light angle in under the dehydrated and shrinking canopy of the much taller planted trees lining the road.

Thus the trial unintentionally demonstrated the negative impact of weed removal, prior to or during meteorological drought, from an area of deep soil moisture deficit, despite moderately dense tree canopy, normal tree leaf turgor, and ground cover and soil surface intermittently wet from light to moderate rain.

On streambanks and in [previously deeply shaded streamside forest, Tradescantia regrowth was much less than expected. In most areas of intervention, follow-up removal was contraindicated by drought.

Having observed these effects, and learned more of the recent and current climate change in Auckland, we now recognise the need to assess deep soil moisture, (noting we lack the expertise or equipment to do this), before planning or undertaking weed control or inorganic refuse removal.

We also suspect there has been a long-term (10-60 years?) increasing desiccation of the roadside bank.
Causes may include:

  • artificial contouring with introduced porous materials during construction
  • absence of trees on the roadside bank until the 1999 planting of the upper roadside (but not the bank below)
  • guttering and piping directly into the stream of all rainwater from the road and the land across the road
  • thinning of canopy due to tree deaths (tanekaha, ti kouka, kanuka, and some unidentifiable) from 1999 (or earlier) to 2019

In the current climate of prolonged drought with short winters and long hot summers, weed control, and any other intervention affecting shade or ground cover in the vicinity of vegetation, needs to be even more restrained and gradual, and the ecological services currently provided by weeds and refuse considered even more carefully before intervention. Some of the locally wild plant species may no longer be viable, especially if planted.

Unable to directly correlate species, tree size, site conditions, run-off patterns and interventions with all observed incidences of severe drought stress, we have been trying to determine the factors influencing each tree’s response to desiccation.

4: Trialling interventions to reduce ongoing desertification of local environments

Since January 2019 we have increasingly focused our Gahnia Grove trial on ways of maintaining or replacing shade and ground cover to allow safe removal of weeds presenting an actual or imminent threat to existing habitat. When time permits we hope to discuss these issues further, along with both trialled and proposed interventions. As always, we welcome any suggestions and experiences, in our Gahnia Grove Trial reports.

Posted by kaipatiki_naturewatch kaipatiki_naturewatch, August 24, 2021 23:52


No comments yet.

Add a Comment

Sign In or Sign Up to add comments