Key Findings of the Banksia Woodland Restoration Project

Mark Brundrett, Margaret Collins, Anna Wisolith and Karen Clarke

Department of Biodiversity, Conservation and Attractions &

The University of Western Australia

1. Introduction to Project and Sites

Objectives of the Banksia Woodland Restoration Project (BWR)

1. Restore and manage banksia woodland

2. Select sites using rigorous ranking process

3. Use scientific approaches to maximise returns

4. Evaluate relative effectiveness of methods

5. Develop monitoring protocols

6. Support community groups and land managers

7. Collate and share information

Commonwealth Government offset for land clearing at Jandakot Airport

Banksia Woodland Monitoring Sites

• 5 Locations in Jandakot and Beeliar Regional Parks

• 31 quadrats for weed control study

• Close to Restoration sites

HarrisdaleSwamp

Rose Shanks

Shirley BallaSwamp

KogolupLake

The Spectacles

Banksia Woodland Monitoring Sites

• 5 Locations in Jandakot and Beeliar Regional Parks

• 31 quadrats for weed control study

• 1 site impacted by fire in 2014

HarrisdaleSwamp

Rose Shanks

Shirley BallaSwamp

KogolupLake

The Spectacles

Weed and Fire Studies

0

5

10

15

20

25

30

35

40

2013 2015 2017 2013 2015 2017 2013 2015 2017 2013 2015 2017

Harrisdale Kogolup Rose Shanks Spectacles

Folia

ge c

ove

r (%

)

Veldt All Dead Veldt Dead on Live

Veldt Live Shrub

Perennial Herb Geophyte

Annual Herb Annual Grass

B Weeds0

5

10

15

20

25

30

35

40

45

50

Folia

ge c

ove

r (%

)

Tree Shrub

Perennial Herb Perennial Grass

Perennial Sedge Geophyte

Annual Herb Annual Grass

Annual Sedge

A Natives

0

20

40

60

80

100

120

0 4 8 12 16 20

Nu

mb

er o

f sp

ecie

s

Month

Seed

Roots

Tuber

Rhizome

Stem

Base

Reseeding

Resprouting

0

5

10

15

20

25

30

35

40

45

0-5 cm 5-10 cm 10-15 cm 15-20 cm 20-25 cm 25-30 cm >30 cm

Ban

ksia

tre

es (

%)

DBH

Resprouting from canopy

Resprouting from trunk

Resprouting from base

Died after resprouting

Died in fire

Weed Study 24 plots Jandakot & Beeliar RPs Fire study Shirley Balla Swamp

Community Grants

• Grants program with $300,000 of funding for banksia woodland restoration by community groups and local governments at 20 locations across the Perth metropolitan area.

Topsoil Transfer Sites

Forrestdale Lake (8 ha)

Anketell Road (40 ha)

Jankakot AirportSite

selection rankings

Restoration Sites

• Sites were weeds-capes before works commenced

• Massive soil seed bank for weeds

• Weedy topsoil was scraped off

2. Reference Site Data for Setting Completion Criteria

• Jandakot Airport Quadrats and Transects

• Tree positions (DGPS) and crown widths on 150 m transects (4)

• Cover and abundance of all spp. from 10x10m plots (12)

• Tree data from 25x25m plots (12)

Relative Importance of Plants

0

200

400

600

800

1000

1200

1400

De

nsi

ty

0

1

2

3

4

5

Pa

ters

on

ia o

ccid

enta

lis

Lyg

inia

ba

rba

ta

Hib

ber

tia

hyp

eric

oid

es

Des

mo

cla

du

s fl

exu

osu

s

Da

syp

og

on

bro

mel

iifo

liou

s

Sch

oen

us

efo

liatu

s

Ph

leb

oca

rya

cili

ata

Da

mp

iera

lin

eari

s

Stir

ling

ia la

tifo

lia

Ca

lytr

ix f

lave

scen

s

Leu

cop

og

on

co

no

step

hio

ides

Pet

rop

hile

lin

eari

s

Styl

idiu

m r

epen

s

Lom

an

dra

her

ma

ph

rod

ita

Co

no

step

hiu

m p

reis

sii

Sch

oen

us

curv

ifo

lius

Hen

sma

nia

tu

rbin

ata

Lep

ido

sper

ma

sq

ua

ma

tum

Lom

an

dra

ca

esp

ito

sa

Co

no

step

hiu

m p

end

ulu

m

Ph

leb

oca

rya

fili

folia

Des

mo

cla

du

s fa

scic

ula

tus

Thys

an

otu

s tr

ian

dru

s

Per

soo

nia

sa

cca

ta

Styl

idiu

m p

ilife

rum

Cro

nin

ia k

ing

ian

a

Arn

ocr

inu

m p

reis

sii

Ca

lect

asi

a n

arr

ag

ara

Ha

emo

do

rum

sp

ica

tum

Laxm

an

nia

sq

ua

rro

sa

Leu

cop

og

on

sp

.

Lom

an

dra

mic

ran

tha

Lom

an

dra

pre

issi

i

Lom

an

dra

su

ave

ola

ns

Thys

an

otu

s th

yrso

ideu

s

Co

ver

Some dominant plants “rescued” for clonal propagation by a specialist nursery

Important Understory Species

Species Rank

Patersonia occidentalis 1

Eremaea pauciflora 2

Lyginia barbata 3

Hibbertia hypericoides 4Amphipogon turbinatus 5

Desmocladus flexuosus 6

Scholtzia involucrata 7Dasypogon bromeliifolious 8

Schoenus efoliatus 9

Phlebocarya ciliata 10

Rank importance values were used to assess the relative dominance of understory plants

Measuring Tree Dominance

• High spatial variability (>0.5 ha survey area needed)

• Over 400 trees / ha in total

• Over 300 banksia trees / ha (main food for Carnaby’s cockatoos)

0

20

40

60

80

100

120

140

160

180

200

De

nsi

ty (

ste

ms

/ h

a)

Use of Reference Site Data in Restoration

Topsoil reference site

Local reference sites

Plant species lists with importance rankings

Existing plot data

Seed collection

targets lists

Direct seeding mix composition

Completion Criteria

Species lists for nursery

orders

Setting Completion Criteria

Criteria Targets

Overall species richness Maximise species richness (not measurable)

Species richness per quadrant Return >60% of native species (10x10m equivalent)

Tree diversity Return all species of trees

Tree density 300 stems/ha for all trees (varies with habitat)

Carnaby’s cockatoo food plants 250 stems/ha for Banksias (in appropriate habitats)

Understorey density Establish 7000 stems/ha (>50% of reference site)

Understorey diversity Return >60% of species

Key understorey plants Monitor 10 most important species separately

*Numeric examples are for Banksia Woodland Restoration Project sites and will need to me adjusted for other habitat types.

3. Comparing Methods for Restoration

Main Restoration Sites

Anketell Road in Jandakot Regional Park

Forrestdale Lake

11.5 ha of Topsoil Transfer in April 2012

Planting

• Over 46,000 plants of 70 species planted

• 3,500 plants of 24 species from cuttings

• Collaborations with Birdlife Austral, Friends or Forrestdale, etc.

Anketell Road

Planted: 32 hectares

Direct Seeding

• Machine seeding was used for broad scale restoration

• Hand seeding was used in areas inaccessible by vehicle

• From 2012-2016 16.5 hectares were direct seeded using 45 kg of seed from 80 different species

Machine seeding

Hand seeding

Anketell Road

Direct seeded: 15.5 hectares

4. Monitoring

• Completion targets (criteria) are based on diversity, density and cover information from reference sites

• Jandakot Airport topsoil source area is the primary reference site, but but some data from local plots was also used

• Monitoring restored areas uses fixed 1x1, 5x5 or 25 x 25 m plots

• Plant diversity, density and cover compared to reference sites quadrat areas of 10 m2 (using 4 adjacent 5 x5 plots)

• The reference site may be atypical (long unburnt, high grazing system)

Adaptive Management Requires Comprehensive Monitoring Data

Reference area species lists with importance rankings

Seed collection targets

Completion Criteria

Species lists for nursery orders and/or direct seeding

Monitoring data quantifying species establishment

Seed collection outcomes

Germination and survival

Plant regeneration from topsoil

Monitoring

5x5 metre quadrats

Total Species Richness

Target Status in late 2015

Maximise native species richness(78 species were present in 12 reference quadrats)

161 native species in total• Highly variable spatially

• Many singletons in plots

Monitoring Plant Density, Cover & Diversity

• Annuals and weeds in 1x1m plots

• Perennial cover in 5x5m plots

• Trees in 25x225m plots

0

2

4

6

8

10

12

14

16

2014Autumn

2014Spring

2015Autumn

2015Spring

2016Autumn

2016Spring

Folia

ge c

ove

r (%

)

Species Area Relationships for Monitoring – Common Plants

• Detection of all plants requires plots to cover the entire area

0

20

40

60

80

100

120

0 2000 4000 6000 8000

Spe

cies

de

tect

ed

(av

era

ge)

Plot area (m2)

All natives

Top 75

Top 60

Top 40

Target

y = 16.84ln(x) - 56.925R² = 0.9671

0

20

40

60

80

100

120

140

160

1 10 100 1,000 10,000 100,000 1,000,000Sp

ecie

s d

etec

ted

Plot size Log m2

Area sampled

Seedling Density in 1 m2 plots• Plot with 39 seedlings of perennial natives (mean = 12, range = 0-42)• Annual native average = 49 (0-163)

Most germinated seedlings died in 2014/15 in driest summer on record

Plant Density

• Stems per ha from 144 - 5 x 5 m quadrats

0

10000

20000

30000

40000

50000

2014Autumn

2014Spring

2015Autumn

2015Spring

2016Autumn

2016Spring

2017Autumn

2017Spring

Pe

ren

nia

l nat

ive

de

nsi

ty

(ste

ms

pe

r h

a)

Anketell Road (topsoil)

Anketell Road (without topsoil)

Forrestdale Lake (topsoil)

Target

Plant Density from Topsoil and Planting

0

2000

4000

6000

8000

10000

12000

2014-A 2014-Sp 2015-A 2015-Sp 2016-A

De

nsi

ty (

ste

ms/

ha)

85 Species

Density of Planted Species Only

0

500

1000

2014-A 2014-Sp 2015-A 2015-Sp 2016-A

Des

nit

y (s

tem

s /

ha)

Planted only (21 species)

Xanthorrhoea preissii Melaleuca thymoides Melaleuca seriata Lepidosperma squamatum SL

Kunzea glabrescens Hypocalymma robustum Hypocalymma angustifolium Eucalyptus todtiana

Eucalyptus marginata Eremaea pauciflora Eremaea asterocarpa Calytrix fraseri

Banksia menziesii Banksia ilicifolia Banksia attenuata Anigozanthos manglesii

Anigozanthos humilis Allocasuarina fraseriana/humilis Acacia saligna Acacia rostellifera

Acacia pulchella

A Complementary Approach

• Topsoil provided high species richness, plant density and many recalcitrant species

• Topsoil alone lacks overstorey species and others with canopy seed

• Direct seeding and planting alone often lacks smaller herbs and annuals

• Thus a combination of methods (topsoil, direct seeding, planting) is most effective

Foliage cover5x5 metre quadrats

0

5

10

15

20

25

2014Autumn

2014Spring

2015Autumn

2015Spring

2016Autumn

2016Spring

2017Autumn

Folia

ge c

ove

r (%

)

Trees

Shrubs

Herbs, sedges and grasses

Restoration Outcomes: Dominant Species

Graph for top 10 native plants by cover. This is dominated by Jacksonia furcellata and Adenanthos cygnorum – disturbance opportunists from topsoil.

Annual plant cover is initially dominated by weeds and Podotheca gnaphalioides, a native disturbance opportunist.

Podotheca gnaphalioides

Jacksonia furcellata

0

500

1000

1500

2000

2500

3000

2014-A 2014-Sp 2015-A 2015-Sp 2016-A

Pla

nt

de

nsi

ty (

ste

ms

/ h

a)

Laxmannia squarrosa

Gompholobiumtomentosum

Scholtzia involucrata

Hibbertia subvaginata

Leucopogonconostephioides

Chamaescilla corymbosa

Jacksonia furcellata

Burchardia congesta

Lechenaultia floribunda

Laxmannia ramosa

Tree Density and Growth

0

50

100

150

200

250

300

350

400

2014 2015 2017

Tree

den

sity

(s

tem

s p

er h

a)

Banksia attenuata Banksia ilicifolia

Banksia menziesii Eucalyptus marginata

Eucalyptus todtiana Other species

0

0.5

1

1.5

2

2.5

2014Autumn

2014Spring

2015Autumn

2015Spring

2016Autumn

2016Spring

2017Autumn

2017Spring

Hei

ght

(met

res)

Banksia attenuata

Banksia menziesii

Eucalyptus marginata

Eucalyptus todtiana

Restoration Outcomes: Perrenial Weeds

Some perennial weeds are common but have low cover (all < 1%).

Annual weeds are more frequent and have higher cover.Anketell Road in 56 - 5 x 5 m quadrats in spring 2015. Red dots show total cover (scale on right).

0

0.2

0.4

0.6

0.8

1

0

20

40

60

80

100

Co

ver

(%)

Fre

qu

en

cy (

%)

Frequency (%)

Cover (%)

Comparing Restoration Potential of Species

Seed quality and germinability determine nursery outcomes

Successful clonal propagation of sedges

Restoration Potential Summary

• Plant diversity is similar (in this case) but vegetation structure differs initially in young banksia woodland on restoration sites

• Long-term trends?

0

10

20

30

40

50

60

70

80

90

0 10 20 30 40 50

Re

fere

nce

sit

e

Restored area

Rank Importance Value Comparison

0 5 10 15 20 25 30

Absent/rare restore

Less restore

Similar

More restore

Only restore

Common Native Plants in Revegetation

• Disturbance opportunists

• From topsoil seed bank

• Decreasing dominance?

• Abundant flowers and seeds

0 5 10 15 20 25 30

Absent/rare restore

Less restore

Similar

More restore

Only restore

Number of Species

Uncommon Native Plants in Revegetation

0 5 10 15 20 25 30

Absent/rare restore

Less restore

Similar

More restore

Only restore

Number of Species

• Low seed or germination

• Resprouting species

• Increasing dominance?

• Clonal Propagation

Restoration Outcomes: Plant Sources

• 160 native plant species by year 4

• Continuum from opportunist to recalcitrant species

113

36

69

22

0

20

40

60

80

100

120

Topsoil Direct seeded Planted Local Opportunistic

Spe

cie

s

Seed Germination and Survival

• Banksia seedling mortality is often high due to invertebrate grazing in winter and severe drought in summer

• Seedlings of most other plants are more resilient than banksias 0%

5%

10%

15%

20%

Direct seeded Cell Nursery Super

Nursery tubestock

Surv

ival

ove

r su

mm

er

Tree Seedling Mortality Rates

• Direct Seeded Banksia attenuata from a 2016 seed germination trial. Deaths from June to October are mostly due to snail and insect grazing and deaths occurring from October to March are due to drought stress.

0

5

10

15

20

25

30

35

40

See

dlin

g d

eat

hs

(% o

f m

ax. e

me

rge

nce

)

Control May

Pre-soaked May

Control June

Pre-soaked June

Winter attrition

Drought deaths

Who is Eating My Seedlings?

• Visible threats to seedlings

0

5

10

15

20

25

30

35

40

1 2 3 4

Caterpillars

KangarooDung

Millipedes

Large Snails

Small Snails

Fenced

Un

fen

ced

The Challenge of Survival

Rainfall deficits 2012-2015

• Several years of planting and seeding are now required to spread the risk of high summer mortality

0

20

40

60

80

100

120

140

160

180

200

Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec

Rai

nfa

ll (m

m)

LTA

2012

2013

2014

2015

2016

Sudden drought in late spring

Outcomes Based on Completion CriteriaCriteria Target

Anketell Roadtopsoil

Forrestdale Laketopsoil

Anketell Roadno topsoil

Total native species richness Maximise native species richness

Total of 160-162 native spp. (highly variable spatially)

71 native spp.

Average native perennial species richness per 10x10m

Return 60% reference quadrats (19 spp.)

30 20 11

Tree diversityPresence of all trees at reference plots

All present

Tree density (stems / ha) >300 243 183 362

Carnaby’s cockatoo food plants (banksias) /ha

250 152 167 222

Average understory species richness per 10x10m

Return 60% of average in reference quadrats (17 spp.)

28(17 – 42)

18(12 – 25)

9(1-18)

Total density of native perennials per 10x10m

Establish 7,000 stems per ha

20,107 6,950 2,675

Annual native plantsNo target set (density is very much lower in reference sites)

17 species4% foliage cover

15 species2% foliage cover

10 species6% foliage cover

Key understory speciesTop 10 most important species from reference plots

All are present, most are common

Not relevant

Weed coverManage serious weeds, especially perennials, monitor annuals

Perennials: 1%Annuals: 4%

Perennials: 1%Annuals: 8%

Perennials: 5%Annuals: 6%

Variability within Sites

Highest native cover: 57% Lowest native cover: <1%

Two quadrats in the same part of Anketell road

Density & Diversity Variations

Density is Lower in areas without topsoil

Diversity is Lower in areas without topsoil

Measuring Area Reaching Targets

• Soil quality issues with Forrestdale Lake limit possible outcomes (buildings and historic land use)

0

10

20

30

40

50

60

70

80

2014Autumn

2014Spring

2015Autumn

2015Spring

2016Autumn

2016Spring

2017Autumn

2017Spring

Pe

rce

nta

ge o

f q

uad

rats

at

targ

et

7,0

00

ste

ms

pe

r h

a

Anketell Road

Forrestdale Lake

Photo Monitoring 1

Spring 2014

Photo Monitoring 2

Spring 2015

Photo Monitoring 3

Spring 2016

Monitoring Total Cover

Ricky Van Dongen & Bart Huntley GIS Branch, DBCA (Anketell Rd)

Downward photos processed for total cover by computer algorithm (eCognition)

5. Sustainable Restoration

Long-term Sustainability & Ecological Functions

Second generation seedlings

Fauna

Pollination

Row 1 - native bees

Row 2 – wasps

Row 3 - flies and butterflies

Row 4 -beetles

Flowering and Seed Set

0.00

0.50

1.00

1.50

2.00

2.50

3.00

3.50

4.00

4.50

5.00

0 20 40 60 80 100 120 140

Can

op

y vo

lum

e (

m3

)Basal diameter (mm)

B. attenuata

B. menziesii

Flowering plants

0

20

40

60

80

100

120

140

160

180

2012 2013 2014 2015 2016 2017

Nu

mb

er o

f sp

ecie

s

Flowering

Total

All plantsBanksias

Harmful Ecological Interactions

• Grazing, weed dispersal and parasites of native plants

Pigface is often spread by rabbits and kangaroos

Galls were common on acacias by year 4

Kangaroo grazing was severe in unfenced areas

Summer Drought & Turnover of Species

• Some disturbance opportunists declining by year 5

• More susceptible to severe summer drought

Disturbance Opportunists

• Common plants in restored areas that were uncommon or not detected at Jandakot Airport (a long unburnt site)

Transform: Square root

Resemblance: S17 Bray Curtis similarity

Date12014_Autumn

2014_Spring

2015_Autumn

2015_Spring

2016_Autumn

2016_Spring

2017_Autumn

2017_Spring

2011_Spring

2D Stress: 0.19

Floristic Recovery Trends

• MDS plots of relative dominance using plant cover

No topsoil

Topsoil sourcereference site(Jandakot Airport)Restoration

areas with topsoil

Conclusions I

1. Grazing and control was essential and weed control very important

2. Effective seed quality control and germination treatments were important

3. Comprehensive reference site and monitoring data was required to guide restoration

4. Restoration methods had a major impact on plant cover and diversity and produced different plant community types

5. Banksia trees had higher mortality than most other species

Conclusions II

6. Rapid tree growth, flowering, seed-set and second generation seedlings have been observed

7. Ecological processes such as turnover of species and biological interactions like flowering and pollination are important and need to be studied more

8. Recovery trends vary due to plant functional groups

9. Recovery trends and timing can possibly be extrapolated but long-term monitoring is required

10. Information required for effective restoration of key species needs to be collected and shared

11. Major problems with offset regulation and management occurred during the BWR project

6. Knowledge Gaps & Research Opportunities

Most common request from community groups (Feb 2016):

• Share results and data to help manage native plants in restoration sites.

Data Required for Effective Restoration

• Species lists for restoration sites with:

1. Habitat type (e.g. upland)

2. Soil/hydrology preferences

3. Relative importance of spp.

4. Seed collection/germination issues

5. Flowering and seed dispersal times

6. Restoration potential and issues

7. Animal associations (e.g. Carnaby’s cockatoo food plants)

8. Fire recovery type and frequency

9. Resilience to weeds and other disturbances

10. Susceptibility to pests and diseases

Revegetation Report Card Banksia attenuataFactor Data

Seed Collection 0-101 seeds per tree (average 15), Jan-Feb

Seed preparation Difficult (heat required to open cones)

Seed storage Requires low humidity and temperature

Seed germination High (30-95%), Inhibited by high temperature

Topsoil seed bank No

Direct seeding results Moderate (0-400 seedlings/ha)

Seedling survival (severe summer drought) Low in first year (3-20%)

Tubestock survival (severe drought) Low in first year (10-20%)

Grazing susceptibility High (kangaroos, rabbits, invertebrates)

Growth rates Fast (0.3-1 m/year height)

First flowering > 5 years

Other impacts on seedlings Grazing, parasitic galls, weeds

Restoration Realities

1. Seed collection 2. Propagation 3. Survival

1. Canopy stored seed 1. Germination easy 1. Weed competition

2. Seed easy to collect 2. Seed treatments 2. Poor soil quality

3. Seed collectible in

sufficient numbers

3. Germination slow

and/or erratic

3. Grazing by

invertebrates

4. Seed quality poor 4. Seed viability low 4. Diseases and pests

5. Seed very hard to get 5. Clonal division only5. Grazing by large

animals

6. Viable seed almost

impossible to get

6. Almost impossible to

propagate6. Severe drought

Species related (1, 2) and site specific (3) factors ranked in according to their importance in banksia woodland restoration