Ecological Archives E086-066-A3

Jonathan R. Rhodes, Clive A. McAlpine, Daniel Lunney, and Hugh P. Possingham. 2005. A spatially explicit habitat selection model incorporating home range behavior. Ecology 86:1199–1205.

Appendix C. Study area, habitat mapping, and radio-tracking data.

 Study area

The study area was the Tomago Sandbeds region of the Port Stephens Local Government Area, New South Wales (Fig. C1). Port Stephens is situated about 150 km north of Sydney on the New South Wales central coast and contains one of the most significant koala populations in the state (Reed et al. 1990). Since European settlement, habitat loss and fragmentation has occurred in the area due to agriculture, urbanization and sand mining (Knott et al. 1998). As described by Lunney et al. (2004), the most common vegetation communities relevant to koalas are melaleuca swamp forest, dominated by broad-leafed paperbark (Melaleuca quinquenervia) and swamp mahogany (Eucalyptus robusta) and forest associations of blackbutt (E. pilularis), smooth-barked apple (Angophora costata), old man banksia (Banksia serrata) and red bloodwood (Corymbia gummifera). Associations of drooping red gum (E. parramattensis) also feature in low open forests. Wet and dry heath communities, grasslands and sedges are also present. Eucalyptus robusta and E. parramattensis have been identified as the koala's most preferred tree species in the area (Phillips et al. 2000).

 

Habitat mapping

Detailed information exists on the distribution of koala habitat in Port Stephens, derived from models of tree species preferences and community-based surveys (Lunney et al. 1998, Phillips et al. 2000). One of the habitat models that Lunney et al. (1998) developed was based upon a fecal pellet survey. This model classified vegetation associations into three koala habitat categories: (1) primary habitat, (2) secondary habitat and (3) marginal habitat. These categories were based on the proportion of preferred tree species in each vegetation association and soil type. We combined a detailed vegetation map of the Tomago Sandbeds area of Port Stephens (D. Lunney, unpublished data) with the vegetation map for the whole of Port Stephens used by Lunney et al. (1998). The two vegetation maps were combined as ArcGIS shape files so that, in areas where we had detailed mapping this replaced the original mapping, but the original mapping was retained elsewhere. Then we classified the vegetation associations as koala habitat based upon the fecal pellet survey habitat model (Fig. C1). This was converted to a raster grid with a 50 × 50 m cell size with the value of each cell representing the most common habitat type in that cell. Few (26%) of one day koala movement distances were less than 50 m and the 50 m resolution was chosen so as to scale the landscape model relative to the lower end of the movement distances. The final habitat map contained seven habitat/vegetation categories: (1) primary habitat, (2) secondary habitat, (3) marginal habitat, (4) sand mining revegetation, (5) cleared land, (6) other vegetation not classified as koala habitat and (7) unmapped/water bodies (Fig. C1). Sand mining revegetation represented those areas which had been revegetated after sand mining operations and which we were as yet unable to classify into a category of koala habitat. Areas classified as unmapped/water bodies were considered unavailable to koalas in the analysis.

 

Radio-tracking data

Between March 1994 and February 1997, koalas from the Tomago Sandbeds region of Port Stephens were radio-tracked following capture or rehabilitation from fire injuries (D. Lunney, unpublished data; Lunney et al. 2004). Prior to release, koalas were marked with a microchip and ear tag and fitted with a collar-mounted radio transmitter. Individuals were then typically radio-tracked for five days per week, but generally only once a week after September 1996. Each day, tracking was conducted until individuals were located visually and then their location were recorded using a global positioning system (GPS).

 

 
   FIG. C1. The Tomago Sandbeds region of the Port Stephens Local Government Area, showing its location in Australia and the distribution of koala habitat. Habitat distributions were derived from the fecal pellet survey based habitat model described by Lunney et al. (1998). The Tomago Sandbeds map is displayed in AGD 1966 zone 56 projected coordinates.

 

We used data from 45 of the individuals (24 female and 21 male), but only extracted successive radio-tracking locations separated by one day for analysis (Table C1). Additionally, we only used successive locations which were separated by less than or equal to the 99thpercentile of the one day movement distances. This was the approach used by Arthur et al. (1996), where the 99th percentile movement distance was used to define the availability radius and movement distances longer than this were not considered. Using a subset of the data in this way was required for the availability radius model and the same subset of data was also used to fit the other models and ensure consistent data between alternative models. Data consistency is necessary to allow valid Akaike’s information criteria (AIC) comparisons between models (Burnham and Anderson 2002). The 99th percentile one day movement distances were 1567 m for females and 1709 m for males. The data subset used for analyses consisted of 2140 pairs of female and 1516 pairs of male radio-tracking locations (Table C1). A small increment (0.1 m) was added to each of the movement distances when fitting the models. This was to avoid the problem of taking the logarithm of zero when the recorded movement distance was zero.

TABLE C1. Summary of the koala radio-tracking data.

Koala

Sex

Radio-tracking start date

Radio-tracking end date

No. pairs of observations

Adriana

Female

21-Nov-94

01-Mar-95

12

Alison

Female

25-Sep-94

28-Feb-95

46

Alison

Female

14-Jun-95

20-Feb-97

153

Coolah

Female

29-Jul-94

13-Feb-96

179

Corrine

Female

25-Sep-94

26-Jan-95

20

Gai

Female

17-Oct-94

27-Dec-94

23

Janie

Female

15-Apr-94

20-Feb-97

253

Johnno

Female

13-Nov-95

09-Feb-96

12

Juanita

Female

17-Jun-94

16-Nov-94

64

Judith

Female

13-Mar-94

23-Jul-96

170

Julia

Female

12-Sep-95

05-Feb-97

73

Kiama

Female

02-Aug-95

19-Feb-97

144

Killibinbin

Female

19-Jun-94

05-Feb-97

139

Lily

Female

16-Nov-95

08-Mar-96

37

Madi

Female

14-Sep-95

08-Dec-96

30

Maggie

Female

01-Aug-95

15-Nov-95

36

Megalong

Female

27-May-94

13-Jan-95

92

Nebby

Female

11-Mar-94

24-Mar-94

3

Nemali

Female

03-Oct-94

24-Feb-97

80

Pinky

Female

10-Mar-94

20-Feb-97

102

Rebecca

Female

26-Sep-94

24-Feb-97

94

Rowena

Female

17-Jun-94

30-Dec-94

71

Sharon

Female

12-Aug-94

24-Feb-97

258

Revlon

Female

29-Jan-96

11-Jun-96

27

Ruby

Female

22-Jul-96

10-Sep-96

22

Alistair

Male

06-Sep-95

03-Jun-96

98

Alistair

Male

21-Oct-96

19-Feb-97

2

Arnie

Male

05-Dec-94

17-Jan-95

15

Bradley

Male

01-Apr-94

25-May-95

124

Copper

Male

27-Oct-94

15-Mar-96

40

Curly

Male

22-May-94

28-Sep-94

31

Duke

Male

01-Feb-95

10-Nov-95

37

Gary

Male

03-Oct-94

26-Jul-95

92

J. B.

Male

16-Jun-94

27-Oct-94

43

Kevin

Male

03-Sep-95

02-Aug-96

48

K. J.

Male

12-Dec-94

21-Jul-95

70

Oliver

Male

17-Jul-94

03-Apr-95

91

Paulie

Male

13-Jul-94

19-Feb-97

262

Rocky

Male

27-Sep-94

20-Feb-97

234

Ryan

Male

08-Jun-94

17-Jun-94

3

Spirit

Male

01-Jun-94

30-Oct-96

209

Toby

Male

09-Feb-95

15-Mar-95

10

Bazza

Male

18-Apr-96

30-May-96

3

Lil Man

Male

23-May-96

09-Dec-96

14

Tim Tam

Male

19-Apr-96

06-Feb-97

48

Toohey

Male

25-Jun-96

05-Feb-97

40

Jamie

Male

26-Sep-96

19-Feb-97

2

† This is the number of pairs of  successive radio-tracking fixes separated by one day and distance apart not greater than the 99th percentile daily movement distance.

‡ These individuals were radio-tracked for two separate periods and were considered independent observations in the analysis.

 

LITERATURE CITED

Arthur, S. M., B. F. J. Manly, L. L. McDonald, and G. W. Garner. 1996. Assessing habitat selection when availability changes. Ecology 77:215–227.

Burnham, K. P., and D. R. Anderson. 2002. Model selection and multimodel inference: a practical information-theoretic approach, Second edition. Springer-Verlag, New York, USA.

Knott, T., D. Lunney, D. Coburn, and J. Callaghan. 1998. An ecological history of koala habitat in Port Stephens shire and the Lower Hunter on the central coast of New South Wales, 1801–1998. Pacific Conservation Biology 4:354–368.

Lunney, D., S. M. Gresser, P. S. Mahon, and A. Matthews.2004. Post-fire survival and reproduction of rehabilitated and unburnt koalas. Biological Conservation 120:567–575.

Lunney, D., S. Phillips, J. Callaghan, and D. Coburn. 1998. Determining the distribution of koala habitat across a shire as a basis for conservation: a case study from Port Stephens, New South Wales. Pacific Conservation Biology 4:186–196.

Phillips, S., J. Callaghan, and V. Thompson. 2000. The tree species preferences of koalas (Phascolarctos cinereus) inhabiting forest and woodland communities on Quaternary deposits in the Port Stephens area, New South Wales. Wildlife Research 27:1–10.

Reed, P. C., D. Lunney, and P. Walker. 1990. A 1986-1987 survey of the koala Phascolarctos cinereus (Goldfuss) in New South Wales and an ecological interpretation of its distribution. Pages 55–74 in A. K. Lee, K. A. Handasyde, and G. D. Sanson, editors. Biology of the koala. Surrey Beatty & Sons, Sydney, Australia.



[Back to E086-066]