internal report 437

Koalas and habitat management on

Kangaroo Island

Seminar Seminar presented at Northern Territory University, 24 April 2003, and at Environmental Research Institute of the Supervising Scientist, 29 April 2003

P Bayliss & P Masters

June 2003

supervising scientist

Koalas and habitat management on Kangaroo Island

Seminar presented at Northern Territory University, 24 April 2003, and at Environmental Research Institute of the Supervising Scientist, 29 April 2003

P Bayliss1 & P Masters2

1 Ecological Risk Assessment Environmental Research Institute of the Supervising Scientist GPO Box 461, Darwin NT 0801

2 South Australia Department of the Environment and Heritage 37 Dauncey Street, Kingscote, South Australia 5223

June 2003

Registry File SG2002/0187

ii

Contents

Abstract 1

Powerpoint presentation 2

iii

Powerpoint slides

Koalas & HabitatHabitat Management KangarooKangaroo Island

Peter Bayliss (eriss) & Pip Masters (SA DEH)

Take Home Messages • Wildlife management is about making value judgements

– aesthetic & technical

• Two key knowledge needs

– how people value wildlife & make decisions

– an understanding of ecosystem dynamics

1 SupervisingSupervising Scientist Division

• Two branches

¾ Environmental Research Institute of the Supervising Scientist (ERISS)

– Environmental protection from impacts of U-mining – Ecology & conservation of tropical wetlands

¾ Office of the Supervising Scientist (OSS)

– Supervision, audit & policy w.r.t. U-mining in ARR

SeminarSeminar OutlineOutline • Pest control background – Herbivore eruptions – Managing overabundance – Bioeconomic framework

• Koalas on KI

• Link with possums in NZ

• KI koalas – where to next?

2 BACKGROUNDBACKGROUND

PestPest ControlControl

Herbivores liberated on islands erupt then crash

Vegetation (food) mirrors reciprocal trend

10

8 Herbivores • Moose Ilse Royal 6 • Possums NZ 4 • Several ungulate sp NZ 2 Vegetation • Sheep Aust rangelands Plant & herbivore density herbivore & Plant 0 0 102030405060

Time (year)

Are these herbivores “overabundant” – what’s the reference point?

3 What is “Overabundance”

• Defined as too many animals but the rigor ends here • Ge n e ra lly 4 c la s s e s (Ca ug h le y 1981)

CLAS S 1: Animals threaten human life or livelihood

CLASS 2: Animals depress abundance favoured species

CLASS 3: Too many animals for their own good

CLAS S 4: The ecological s ystem is off its equilibrium

• All apply to koalas on Kangaroo Island

Definition of “damage” caused by Class 1 & 2 overabundance

• All animals have an impact on their consumptive resources (food & shelter)

• Damage occurs when impact causes economic or environmental harm

• How one defines “harm” depends on how one makes a living

4 Managing Pest Impacts

z Involves making c hoices

– how much management intervention at what cost ($) ? – what benefit is delivered?

z Challenge is to make choices that are

– sensible – pragmatic – defensible

z Requires benefits & costs to be balanced at least

– past focus on “activity-based” management – new focus on “damage-based” management within a budget – involves complex decision making - use modelling tools

Bioeconomic Modelling

Pest control

Outputs Economic inputs Pest density (additional income or (costs of control) improved public asset)

Damage reduction

Benefit/cost analysis

Monetary benefit/cost analysis Benefit maximisation Cost minimisation

5 There are 3 key sub - models

1. Damage – pest density relationship

2. Exponential cost-of-control curve

3. Population growth response

1. Damage – density relationship

160 140 120 100 Socially acceptable? 80 60 40 Threshold 20 Level of DamageLevel of 0 024681012

Pe st Dens ity

6 2. Cost - of - control curve Choosing the right “Target Density”

Threshold density

Socially acceptable damage-density level TARGET DENSITY

Buffalo – Arnhem Land Bayliss (1985)

3. Population growth response

1.6

1.4

1.2 • Logistic population growth 2 st 1.0 model as 1 approx 0.8 0.6 • Rapid recovery rate for pest 0.4 species 0.2 Density feralkm /

0.0 024681012• Manage control interval (yrs) &

Years initial & maintenance kill rate

1.6 1. 6

1.4 1. 4

1.2 initial control 1. 2 2 2 1.0 1. 0

0.8 0. 8

0.6 0. 6

0.4 0. 4

0.2 0. 2 Density feral / km feral Density Density feral / km maintenance control 0.0 0. 0 024681012 024681012

Years Years

7 Target density & control technology

MORTALITY CONTROL FERTILITY CONTROL + NUMBERS NUMBERS

TI ME TI ME

INTEGRATED CONTROL = = SUSTAINED SUPPRESSION NUMBERS

TI ME

Summary How to manage damage

• Clearly define damage caused by overabundance • Identify the damage - density "threshold" • Identify a socially acceptable level of damage & corresponding animal density • Use bioeconomic & ecological frameworks (models) to guide control program cost - effectively

8 KOALASKOALAS

KANGAROO ISLAND

KANGAROO ISLAND SOUTH AUSTRALIA

• Jewel in the crown of conservation lands in SA • Koala is only introduced major animal pes t species

10 THE STORY

• Koalas introduced KI 1925 in attempt save species thought to be diminishing on mainland (85 released: 1923 – 1964)

THE STORY

• Populations flourished – by 1965 severe over-browsing damage to prefered food trees (e.g. Manna gum – Eucalyptus viminalis)

10 THE MAIN IS S UE S • Koalas quickly defoliate trees – now a “threatening process” to large areas of Eucalypt forests

• High risk of local extinction of preferred species

• Closest rival in reputation is introduced possum in NZ

• 1996 – SA Koala Task Force & National Koala Conservation S trategy established

SA Koala task Force Recommendations (Possingham et al. 1996)

• Do nothing • Protect & restore degraded habitat • S uppress fertility via introduction of Chlamydia virus (suggested by Koala Foundation) • S uppress fertility by surgica l or hormona l methods • Translocate surplus animals to other sites • Culling

11 BUT .... different perceptions of value of koalas

• Koalas - high & favourable profile in the Australia physch • Ob vious tou ris t be n e fits to KI ($$$$) • But cause tree damage – so big conflict in values • Determines how they are managed

Response to SA Koala task Force Recommendations

• Do nothing – not an option • Chlamydia rejected on animal welfare grounds • Adverse national & international reaction to culling – forget totally • No room for translocations on KI, shift some to SA mainland • Maintain low koala densities on KI via fertility control • Implement habitat protection & restoration program on KI • Yes - develop community education program • Yes - expand research & management effort (1996-2000)

12 What happened next ?

• Politics took over – urgent action required

• Management was activity-orientated, not strategic

• e.g. in absence reliable information, arbitrary sterilisation (e.g. 70% pop) & translocation targets were set

Koala Rescue Program • Funds available to research tree damage & manage koala populations on KI by surgical sterilisation & translocation • Called “Koala Rescue” Program not “Habitat Rescue”

13 Cos t to Public ?

• In 1 year alone – 1997/98 FY

• 3,396 sterilised

• 1,105 relocated off Kangaroo Island

• Total cost = $300,00 or $131 / koala

Is the program working?

• Action pla n ba s ed on initia l pop es tima tes for Cygnet R iver catchment (most browsing damage) = 3,000 – 5,000 koa la s

• Other ca tchments not surve yed – whe n include d = 26,000 (i.e. 5 – 9 times extent of perceived problem)

• Rapidly running out of sites on mainland to dump sterilised koalas from KI

• Answer up front: locally yes, globally no – unsustainable

14 Then what happened ?

• P rogra m revie wed 2000 - request to Ma rs upia l CR C to:

– analyse all available information on browse damage (1996 - 2000) – help estimate koala densities island-wide – develop qua ntita tive decis ions s upport tools (bioeconomic models , ecological risk assessment) to guide strategic management

• Until this request no previous a na lysis of informa tion ma de because of pressures of operational actvities

• Results follow

Ma n a g e m e n t D a ta

• R e c e nt (1996 – 2000) • His toric a l (1925 – 1995)

15 Funding allocated to koala management on Kangaroo Island (1996 - 2001)

Management Activity 1996/97 1997/98 1998/99 1999/00 2000/01 TOTAL

Fertility control $190,000 $150,000 $155,000 $170,000 - $665,000 Coordination & Monitoring $75,000 $70,000 $20,000 $30,000 $15,000 $210,000 Translocation $25,000 $50,000 $15,000 --$90,000 Community education $45,000 $30,000 $10,000 --$85,000 Island wide pop estimate - - - - $185,000 $185,000 & program review

TOTAL $335,000 $300,000 $200,000 $200,000 $200,000 $1.235M

Koala population estimate (& SE%) per catchment (# sample sites) on Kangaroo Island (2000)

Catc hment Population Standard Error (number of sites) estimate (%) North Coa s t (25) 5,636 31 S ou th We s t (22) 5,884 30 Eleanor-Timber Ck (24) 5,196 21 Finders Chase NP (21) 2,993 31 Cyg n e t R iv er (27) 5,442 31 TOTAL (119) 25,871 12 P re-1996 e s tim a te 3,000 – 5,000 None

16 Predicted extinction rates of Manna gum trees in management units if no intervention (from 2000)

Management Koala density % Manna gum with Years to all trees Unit Nos / ha <%50 canopy cover severely damaged Flinders Chase 5.01 94 2 Timber-Creek 1.46 75 9 North Coast 1.64 79 7.5 Cygnet River 1.85 60 15 South west - 90 3.5

Browse damage classes & associated koala population estimates (1996-1999)

• 119 sites – mark-recapture “double count” method used in sites (va rying 1-10 ha ) to es tima te koa la dens ity

• Koa las tagged, sexed, aged, weighed, sterilised & released or sterilised & tra nsloca ted

• All species of trees counted & classified according to damage

Class 1: Crown normal (no visible signs) Class 2: Thinning of crown (up to 50% defoliation) Class 3: Crown sparse (50-80% defoliation) Class 4: > 80% defolia tion, predomina nt epicormic growth Class 5: Crown absent, tree dead

17 Scotch Thistle Flat – Flinders Chase Na tiona l P a rk, Ka ng a roo Is la nd (1964 – 2000)

Scotch Thistle Flat - Proportion trees severely defoliated (class 5 candidates, won't recover) 1.0

0.8 P = 0.03T 0.6 (R2 = 0.95, n= 8, P<0.001)

0.4

Proportion (P) Proportion 0.2

0.0 0 5 10 15 20 25 30 35 40 Years since 1963 (T)

• 47 koalas released in 1964 • 35 years to extinction of Manna gum habitat at 3% p.a. tree loss rate

Time trends koala density & severe Damage Classes of Manna gum trees, Cygnet River catchment (1996 – 2000)

Density D1 (healthy trees) vs Time Koala density vs Time 1.8

5.0 1.6 1.4 4.0 1.2

3.0 1.0 0.8 D1 (nos/ha) D1 2.0 0.6

0.4

Koala Density (nos/ha) 1.0 0.2

0.0 0.0 1995 1996 1997 1998 1999 2000 2001 1995 1996 1997 1998 1999 2000 2001

Time (Year) Time (Year)

Density D4 trees (> 80% canopy loss) vs Density D5 trees (dead) vs time Time 2.0 4.0

3.5

3.0 1.5

2.5

2.0 1.0

1.5 D5 (nos/ha) D4 (nos/ha) 1.0 0.5

0.5

0.0 0.0 1995 1996 1997 1998 1999 2000 2001 1995 1996 1997 1998 1999 2000 2001

Time (Year) Time (Year)

18 Time trends of intermediate Damage Clas s es of Manna gum tre e s , Cyg n e t R ive r c a tc h m e n t (1996 – 2000)

Density D2 trees (<25% canopy loss) vs Density D3 trees (~ 50% canopy loss) vs 4.0 Time Time 8.0

3.0 6.0

2.0 4.0 D2 (nos/ha) D3 (nos/ha)

1.0 2.0

0.0 0.0 1995 1996 1997 1998 1999 2000 2001 1995 1996 1997 1998 1999 2000 2001

Time (Year) Time (Year)

As Class 4 trees (80% canopy loss) recovered they enter Class 3 & then 2, which explains their increasing trends

ME S S AGE F R OM D AT A Control program seems to be working in the Cygnet River catchment !

19 Damage – dens ity relations hips & thres holds

Density D4 trees vs Koala density kangaroo Density D4 (> 80% canopy loss) Manna Gum Isld (annual data, all Eucalyptus sp, 1996-2000) Trees vs Koala Density Cygnet River (1996 - 1997) 3.5 20 3.0 D4 = 0.75KD - 0.21 2 16 2.5 R = 0.95, P<0.005 D4 = 1.20KD - 2.05 R2 = 0.75, P<0.01 2.0 TD=0.28 / ha 12

loss) 1.5 TD = 1.7/ha 8 1.0 D4 (nos/ha)

0.5 4 D4 (nos/ha > 80% canopy 80% canopy > (nos/ha D4

0.0 0 0.0 1.0 2.0 3.0 4.0 5.0 0 4 8 12 16 Koala density (nos/ha) Koalas / ha

B ut confounding effect - koala density also a function of tree density (more trees more koala s more damag ed trees ) – us e multivariate damag e rate function or proportion of damaged trees

Manna gum damage rate, koala density & tree dens ity

Damage rate (Classes 3 & 4 > 50% canopy loss) vs Mean Koala density - Manna gum habitat 15 Cygnet River (KI), 1996 - 2000

10

5

0 (nos/ha/yr) 051015 Damage rate trees -5 Koala Density (nos/ha) Trees/ha T12 -10 Damage rate = -1.23 + 0.96Mean KD - 0.26TD T20 2 (R = 73%, n=36, all coeffs sig, reg P<0.001) T5

20 Final model us ed to estimate thres hold damage level & target density for control

% Manna gum trees in D4 damage class vs % koalas, Cygnet River, kangaroo Isld (1996- 25 2000)

20

Damage threshold = 20% 15 Mean density koalas = 4.2/ha (1997) x 0.2

10 Target Density = 0.84 ~ 1.0/ha

%trees in D4 class 5

0 0 20406080100

% of initial control density

Variation in threshold damage density across landscape & habitat

• Island-wide, across all species (TD = 0.3 / ha )

• Manna gum only in Cygnet River catchment (TD = 0.8 / ha )

• Site-specific variation – damage rate increases with decreasing Manna gum tree density (TD = 3.0 – 7.0 / h a )

• Management needs to account for habitat variations in composition of Eucalyptus species & tr e e d e n s i t y

21 Some koala habitat relationships

• GLM used to predict spatially explicit variables (Xs) that may influence patchy distribution & abundance of koalas (Y, density) on KI

• Source data - 119 survey sites across management units (catchments)

• Variables entered in model:

– Management unit of site (catchment) – % composition of all Eucalyptus species

• Model: Koala density = E. viminalis + E. leucoxylon + Management Unit (Total R2 = 57%; Manna gum R2 = 29%)

• Tree density not initially entered but would explain large % residual

Decision support tools bioeconomic model

• Damage-density relationship = target density control

• Cost function ($ / koala) = fixed + variable costs

• Population response model (logistic model as 1st approx)

• Combined models - “What if” scenario simulator

• Simulate combinations of different management options (kill – shoot, euthanasia; translocate; sterilise)

22 Cost functions

Program Management Manager (& on-costs) $48,100 p.a. Equipment basics $10,000 p.a. Office costs $12,000 p.a.

Monitoring Survey team $645 p.day Vehicle $42 p.day 1. F ix e d c os ts Total / day $686 p.day Total 6 vweeks $20,595 p.6 weeks

Others Vet surgical procedure $50 p.koala Translocation flight $45 p.koala Extra handling time KI $5 p.koala Handling time - release $5 p.koala Ammunition (.223) $2.50 p.koala

2. Variable costs – search effort vs density

Average search time (minutes) 1996 - 1998

300

250 ST = 39.2 KD -0.8675 Catching team (3 people) $65 p.hour 200 R2 = 68.3%, R =0.8264 Vehicle cost $5.50 p.hour n=63, P<0.001 150 Vet time lethal injection $110 p.hour Marksman $30 p.hour 100 Marksman assistant $21 p.hour

Search time (ST, mins) 50

0 0.0 2.0 4.0 6.0 8.0 10.0 12.0 14.0 Koala density (KD, ha-1)

Catching team cost $65/hr * hrs/koala @ density + other costs

23 Management scenarios Model inputs for simulation run

1. Initial cull &/or fertility control 1 Year 6 Month 0.8 Proportion culled Logistic population sub-model 10 Years post-initial control 50 Simulation interval (yrs)

2. Maintenance cull control 0.26 = rm p.a. 3.0 = Theta 0.5 Proportion culled 0.00 = imax p.a. 0.33 = Max birth rate p.a. 5 Interval (years) 0.000 = min density / ha 25 Number of cull ops 4.50 = K, carrying capacity / ha 3. Fertility control 0.07 = death rate p.a. in absence of births 0.00 Proportion sterilised 100 Longevity of sterility (years) 5 Interval (years) 20 Number of fertility control ops Manna gum tree damage sub-model

4.Target density 1.00 ha-1 -1 20.5 Mean % of Class 5 dead trees 5. Density floor = 0.00 ha 19.3 Mean % of Class 4 trees (>80% defoliation)

6. Only cull without eartags 0.7 Density (/ha) of Class 5 dead trees nY or N 2.5 Density (/ha) of Class 4 trees (>80% defoliation) 8.7 Mean tree density/ha 7. 1 if euthanasia 0 if shoot = 1 27.0 D4 damage rate (% p.a.) in absence of removal 8. Prop. of 'cull' translocated = 0.00 24.0 D5 damage rate (% p.a.) in absence removal 91Habitat or control site 2000 Control area (ha)

Combining lethal & sterilisation control options to get below target density rapidly

5.0

4.0 Sterile Fertile Total 3.0 Target 80% initial kill 2.0 80% sterilisation at 5 yr Koalas / Koalas ha in te rva ls for 50 yrs 1.0 Total cost = $1.5M 0.0 st 0 5 10 15 20 25 30 35 40 45 50 TD reached 1 year

Years

24 Sterilisation only option

5.0

4.0 Sterile Fertile Total 3.0 Target No initial cull 2.0 80% sterilisation at 5 yr Koalas / ha Koalas in te rva ls for 50 yrs 1.0 Total cost = $1.7M 0.0 0 5 10 15 20 25 30 35 40 45 50 TD re a c h e d 33 yrs

Years (> extinction rates trees)

Model outputs for simulation run Post control efficiency R ec overy of Manna gums 3.08 Average post-control density ha-1 1 Years to 1st reach or pass target density 50 Years above target density 0 Years leas than or equal to target density

100

80 60 Cos t of control 40 20

% of Damage Class 0 Costs/ha Total costs 0 1020304050 He althy tre e s Time (50 Years) Initial cost ($/ha) $43.95 $87,906 Severely damaged Dead tree s An. maintenance cost ($/ha/yr $27.70 $55,398 Total cost ($/ha) for $1,401.21 $2,802,425 50 Years

25 Validation of koala bioeconomic model

• Model used Cygnet River catchment data (1996-1999) then c om pa re d w ith 2001 da ta • Generally predictive (= useful):

– pos t-c on trol koa la den s ity (2.00 v s 1.97 +0.15 / h a ) – % healthly MG trees (73% vs 81%) – % severely damaged MG trees (7% vs 7%) – % de a d MG (20% v s 12%) – Predicted & actual costs about same (~ $600,000)

• B ut what about the other 20,000 koalas in other catchments?

POSSUMSPOSSUMS in New Zealand

26 Comparing koalas on KI with possums in NZ

• P os s um ma x rate of inc rea s e = 0.25 p.a cf Koala = 0.26 p.a. (mainland data, Chlamydia free pop); similar body weights

• Possums have erupted & crashed after 150 yrs; koalas still erupting a fte r 78 yrs

• Small numbers of both released on islands with superabundant food – vacant habitat

• Interactive plant - herbivore model predicts that populations will crash after vegetation crashes

• Main manag ement aim Koala s on KI is to a void veg etation c ras h (& death & starvation of thousands koals) -nip the problem in the bud

What population model do Kiwis use? Same as for koalas

z = 7

z = 3

z = 1 Rate ofincrease Rate (r) Possum density Possum

Population size Time Generalised logistic model

• Assumes vegetative food resources have no dynamics • Hides ecological processes rather than exposes them • Inappropriate for modelling complex ecological processes

29 Comparing koalas on KI with possums in NZ

• P os s um ma x rate of inc rea s e = 0.25 p.a cf Koala = 0.26 p.a. (mainland data, Chlamydia free pop); similar body weights

• Possums have erupted & crashed after 150 yrs; koalas still erupting a fte r 78 yrs

• Small numbers of both released on islands with superabundant food – vacant habitat

• Interactive plant - herbivore model predicts that populations will crash after vegetation crashes

• Main manag ement aim Koala s on KI is to a void veg etation c ras h (& death & starvation of thousands koals) -nip the problem in the bud

What population model do Kiwis use? Same as for koalas

z = 7

z = 3

z = 1 Rate ofincrease Rate (r) Possum density Possum

Population size Time Generalised logistic model

• Assumes vegetative food resources have no dynamics • Hides ecological processes rather than exposes them • Inappropriate for modelling complex ecological processes

28 New Possum – Plant Model

• Hinau (Elaeocarpus dentatus) - endemic NZ ha rdwood, lives to about 400 years

• Possums eat hinau fuit

• Hinau has fruit MAS TING cycle of 2 + years

• Possum rate of increase positively correlated to annual crop of hinau fruit

• S trong interaction - hinau fruit production increases dramatically where possums eradicated

Importance of damage-density thresholds

What target density to restore “masting” cycle in hinau fruit production?

Target = 80% reduction Target = 90% reduction

16.0 16.0

14.0 14.0

12.0 12.0 10.0 10.0 8.0 8.0

6.0 6.0

4.0 4.0 2.0 2.0

POSSUM & HINAU DENSITY HINAU POSSUM & 0.0 DENSITY HINAU POSSUM & 0.0 0 1020304050607080 0 1020304050607080

YEAR YEAR

Possums Hinau

29 Where to next for koalas?

• Temporal dynamics complex – abandon Logistic model & progress to interactive plant - herbivore (koala – tree) model - possible

• Spatial dynamics complex – link temporal dynamics to habitat processes across landscapes (= landscape complementation)

• Koalas on KI are a closed population – so tractable

• Assume increased model realism & predictability leads to increased utility – is this really true?

Interactive koala - tree model

Phase plane trajectory H & V in absence 10 predators 8 10

6 8

6 4 4

2 2 Herbivore density Herbivore Plant & herbivore density herbivore & Plant 0 0 0 102030405060 0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0

Time (year) Plant density

dV  V    − 1Vd Explicit model functions mVVr 1 −=  1 ()1 −− eHc dt  KV  • Plant growth function dH • Functional response herbivores − 2Vd []2 ()1 −+−= ecaH • Numerical response herbivores dt

30 1. Plant growth response

800

400 BIOENERGETICS model has 3 key explicit & linked nonlinear 0 energy transfer functions

Growth (kg/ha/3 mo) -400 0 200 400 600 800 1000

Vegetation (kg/ha)

2. Functional response 3. Numerical response 100 0.5 0.0 80

60 -0.5

40 -1.0

20 -1.5 Food intake (g/day) intake Food 0 p.a.) (r increase of Rate -2.0 0 200 400 600 800 1000 0 200 400 600 800 1000 Food (vegetation, kg/ha) Food (vegetation, kg/ha)

1. PLANT GROWTH FUNCTION

Foliage dynamics with & without browsing (S oo Lim PhD thes is )

Leaf Productivity

0.60 VH

0.50

0.40

0.30 FC VH – Victor Harbour no koalas

summer FC – Flinders Chase, KI, koalas 0.20

0.10

Instant. growth rate over 0.00 020406080100

Starting biomass of leaves (%) relates to tree 100% - Defoliation - 0% damage classes

31 2. F unc tional R es pons e Food intake vs availability (S oo Lim PhD thesis )

500

400

300

200

100 # leaves/koala/day

0 0 500 1000 1500 2000 Number leaves/branch

Need to convert number leaves to biomass (dry weight kg)

3. Numerical response: rate of increase vs food level

• No data on ma x rate inc rea s e (rm ) for KI koalas • No data on rate of inc rea s e over rang e of food a vaila blities • Us e the “ba c k door” a pproa c h to derive num eric a l res pons e function

Biomass conversion principle Rate of increase

Rate of food intake

32 Spatial Dynamics – coarse grain How does “habitat quality” affect interaction between animal populations and their food resources ?

FOOD HABITAT QUALITY AVAILABILITY

FORAGING EFFICIENCY (d1)

− 1Vd e1ccintake food 1 ()−= e1ccintake FOOD OFFTAKE

DEMOGRAPHIC EFFICIENCY (d2)

− 2Vd incre ase ofrate ofrate increase r 2 ()−+−= e1ca

DEMOGRAPHY

Landscape complementation approach

• Shows importance of biomass conversion principle

• Yield insights into consumer-resource interactions, especially between behavioural tradeoffs & habitat quality (e.g. foraging, thermoregulation)

• Implications for metapopulation analysis & linking population processes to landscape scales

• Multiple limiting factors supported

33 Spatial Dynamics – fine grain (impacts of koala browsing on forest dynamics)

• Koalas selectively browse Eucalyptus species according to their relative palatability & availability in the landscape

• Koalas influence dynamics of Eucalyptus forests by selective browsing, which in turn affects species plant growth & survival rates

• Hence, composition of forests inhabited by koalas is influenced by koala browsing

• Which in turn influences koala browsing (negative feedback loop – regulation)

Forest Dynamics & Koala Browsing (species & size composition matrix embeded deep in the food axis)

Species i Species i

Species i Species i j j j j Forest composition Forest composition Intake rate Forage availability t t+1 Iij estimated from t F ΣPVij PVij Tij and aij ΣPVij Size-class Size-class Size-class Size-class Size-class Size-class

Species i Species i j

j Fecundity Defoliation rate t t

φPij Size-class Size-class Size-class Need longitudinal tree Species i

j data c.f. to cros s - Koala density Growth sectional data D t Size-class Size-class

Species i i.e. tag many trees j Survival t Size-class Size-class

34 Wildlife Management Upshot

• Aesthetic values judgments – no right or wrong, just rights • Challenge - resolve conflict between dichotomous views

• Technical value judgments – there is a right & a wrong • Challenge – being right

THE END

35