18
Canadian Journal of Zoology XXXXXX Megan L. Hornseth and Robert Rempel. 201x. Seasonal resource selection of woodland caribou (Rangifer tarandus caribou) across a gradient of anthropogenic disturbance Appendix A. Seasonal Resource Selection for woodland caribou We fitted seasonal resource selection probability functions to categorized habitat use, as determined through seasonal utilization distributions of 240 caribou in northern Ontario. High-use represents ‘selected’, and low-use ‘not- selected’ habitat. Each region (Developed Boreal Shield (DBS), Undeveloped Boreal Shield (UDS), and Hudson Bay Lowlands (HBL)) contained 2 ranges. The models for each hypothesis were developed based on a thorough literature review of habitat selection and use by boreal woodland caribou across Canada (Bradshaw et al. 1995, Racey et al. 1999, Rettie and Messier 2000, Dyer et al. 2002, Vors et al. 2007, Fortin et al. 2008, Bowman et al. 2010, Dussault et al. 2012). Variables are summarized in Table 1. Spring was delineated as April 16 to June 21, summer was from June 22

University of Toronto T-Space€¦ · Web viewCanadian Journal of Zoology XXXXXX Megan L. Hornseth and Robert Rempel. 201x. Seasonal resource selection of woodland caribou (Rangifer

  • Upload
    others

  • View
    0

  • Download
    0

Embed Size (px)

Citation preview

Page 1: University of Toronto T-Space€¦ · Web viewCanadian Journal of Zoology XXXXXX Megan L. Hornseth and Robert Rempel. 201x. Seasonal resource selection of woodland caribou (Rangifer

Canadian Journal of Zoology XXXXXX

Megan L. Hornseth and Robert Rempel. 201x. Seasonal resource selection of

woodland caribou (Rangifer tarandus caribou) across a gradient of anthropogenic

disturbance

Appendix A. Seasonal Resource Selection for woodland caribou

We fitted seasonal resource selection probability functions to categorized habitat use, as

determined through seasonal utilization distributions of 240 caribou in northern Ontario.

High-use represents ‘selected’, and low-use ‘not-selected’ habitat. Each region

(Developed Boreal Shield (DBS), Undeveloped Boreal Shield (UDS), and Hudson Bay

Lowlands (HBL)) contained 2 ranges. The models for each hypothesis were developed

based on a thorough literature review of habitat selection and use by boreal woodland

caribou across Canada (Bradshaw et al. 1995, Racey et al. 1999, Rettie and Messier 2000,

Dyer et al. 2002, Vors et al. 2007, Fortin et al. 2008, Bowman et al. 2010, Dussault et al.

2012). Variables are summarized in Table 1. Spring was delineated as April 16 to June

21, summer was from June 22 to September 21, fall was from September 22 to December

21, and winter was delineated as December 22 to April 15.

Table A1. Comparison of logistic regression models based on the 5 a priori hypotheses

predicting the relative probability of caribou probability of use across 4 seasons and 3

regions. Models are not shown for Pagwachuan summer because the summer model from

Nipigon was evaluated and applied, and for Spirit spring, fall, and winter models because

models from Kinlock performed well. Relative measures of model support are shown,

including, -2* Loglikelihood (-2*LL), number of estimated parameters (k), Akaike’s

Information Criterion (AIC), AIC difference (∆i), AIC weight (wi), receiver-operating

Page 2: University of Toronto T-Space€¦ · Web viewCanadian Journal of Zoology XXXXXX Megan L. Hornseth and Robert Rempel. 201x. Seasonal resource selection of woodland caribou (Rangifer

criterion Area Under the Curve (AUC), specificity and precision. Models with AUC<0.6

were not considered plausible and not included in AIC model selection set. N is the

number of individual caribou used in model development and evaluation.

Page 3: University of Toronto T-Space€¦ · Web viewCanadian Journal of Zoology XXXXXX Megan L. Hornseth and Robert Rempel. 201x. Seasonal resource selection of woodland caribou (Rangifer

DBS - Nipigon Range

Hypothesis -2 *LL k AIC ∆i wi

AUC

Specificity

Sensitivity

Spring (N = 84)

Global Model3189.7

8 113211.7

8 0.00 1.00 0.740.94 0.96

Competition3245.9

6 73259.9

6 48.18 0.00 0.730.94 0.72

Browse Refuge3420.6

6 73434.6

6222.8

8 0.00 0.660.93 0.65

Browse Availability3560.2

2 63572.2

2360.4

4 0.00 0.600.91 0.40

Disturbance - 3 - - - 0.53 0.93 0.80Summer (N = 80)

Global Model3003.2

8 113025.2

8 0.00 1.00 0.770.95 0.75

Browse Refuge3149.5

0 73163.5

0138.2

2 0.00 0.730.96 0.83

Competition3173.8

6 73187.8

6162.5

8 0.00 0.710.96 0.78

Browse3260.3

6 63272.3

6247.0

8 0.00 0.690.95 0.77

Disturbance3421.3

2 33427.3

2402.0

4 0.00 0.600.94 0.83

Fall (N = 70)

Global Model3194.4

8 113214.4

8 0.00 1.00 0.780.93 0.59

Competition3288.7

8 73302.7

8138.2

2 0.00 0.770.94 0.69

Browse Refuge3448.8

8 73462.8

8162.5

8 0.00 0.740.94 0.77

Browse Availability 3581 63593.0

0247.0

8 0.00 0.710.92 0.61

Page 4: University of Toronto T-Space€¦ · Web viewCanadian Journal of Zoology XXXXXX Megan L. Hornseth and Robert Rempel. 201x. Seasonal resource selection of woodland caribou (Rangifer

Disturbance 3840.5 33846.5

0402.0

4 0.00 0.600.90 0.50

Winter (N = 62)

Global Model3841.7

6 113863.7

6 0.00 0.99 0.730.94 0.67

Competition3924.1

4 73938.1

4 74.38 0.01 0.720.93 0.76

Disturbance4115.4

4 34121.4

4257.6

8 0.00 0.640.93 0.80

Browse Refuge4126.4

6 74140.4

6276.7

0 0.00 0.660.91 0.72

Browse Availability4130.1

6 64142.1

6 278.4 0.00 0.650.92 0.79

Page 5: University of Toronto T-Space€¦ · Web viewCanadian Journal of Zoology XXXXXX Megan L. Hornseth and Robert Rempel. 201x. Seasonal resource selection of woodland caribou (Rangifer

DBS - Pagwachuan Range

Hypothesis -2 *LL k AIC ∆AIC WeightAUC

Specificity Precision

Spring (N = 77)

Global Model 3272.2611 3294.26 0.00 1.00 0.78

0.94 0.70

Browse Refuge 3338.96 7 3352.96 58.70 0.00 0.76 0.93 0.67Browse Availability 3366.46 6 3378.46 84.20 0.00 0.75 0.93 0.58Competition 3603.66 7 3617.66 323.40 0.00 0.70 0.94 0.76Disturbance 3721.94 3 3727.94 433.68 0.00 0.64 0.93 0.77Fall (N = 40)

Global Model 3534.911 3554.90 0.00 1.00 0.82

0.95 0.78

Browse Refuge 3739.68 6 3751.68 196.78 0.00 0.79 0.93 0.67Browse Availability 3761.82 5 3771.82 216.92 0.00 0.78 0.93 0.71Competition 3802.68 7 3816.68 261.78 0.00 0.78 0.93 0.77Disturbance 4182.82 3 4188.82 633.92 0.00 0.67 0.89 0.53Winter (N = 44)

Global Model 3409.2411 3431.24 0.00 0.98 0.78

0.96 0.86

Competition 3487.42 7 3501.42 70.18 0.02 0.75 0.97 0.88Browse Refuge 3617.04 7 3631.04 199.80 0.00 0.73 0.94 0.71Browse Availability 3641.16 6 3653.16 221.92 0.00 0.72 0.93 0.70Disturbance 3724.88 3 3730.88 299.64 0.00 0.65 0.99 0.99

Page 6: University of Toronto T-Space€¦ · Web viewCanadian Journal of Zoology XXXXXX Megan L. Hornseth and Robert Rempel. 201x. Seasonal resource selection of woodland caribou (Rangifer

UBS – Kinlock Range

Hypothesis -2 *LL k AIC ∆AIC WeightAUC

Specificity Precision

Spring (N = 117)

Global Model 3444.4611 3466.46 0.00 1.00 0.76

0.94 0.74

Competition 3521.98 7 3535.98 69.52 0.00 0.74 0.94 0.77Browse Refuge 3530.86 7 3544.86 78.40 0.00 0.74 0.93 0.70Browse Availability 3625.96 6 3637.96 171.50 0.00 0.72 0.91 0.73Disturbance 3906.22 3 3912.22 445.76 0.00 0.60 0.94 0.72Summer (N = 111)

Global Model 4496.7011 4518.70 0.00 1.00 0.72

0.88 0.71

Browse Refuge 4572.04 7 4586.04 67.34 0.00 0.71 0.89 0.70Competition 4692.52 7 4706.52 187.82 0.00 0.68 0.89 0.72Browse Availability 4710.10 6 4722.10 203.40 0.00 0.68 0.86 0.70Disturbance 4930.02 3 4936.02 417.32 0.00 0.60 0.90 0.71Fall (N = 107)

Global Model 3720.4211 3742.42 0.00 1.00 0.77

0.91 0.74

Competition 3856.26 7 3870.26 127.84 0.00 0.75 0.92 0.70Browse Refuge 4043.08 7 4057.08 314.66 0.00 0.71 0.90 0.55Browse Availability 4154.48 6 4166.48 424.06 0.00 0.68 0.94 0.91Disturbance 4342.12 3 4348.12 605.70 0.00 0.61 0.91 0.62Winter (N = 89)

Global Model 3547.2011 3569.20 0.00 1.00 0.73

0.93 0.71

Competition 3665.22 7 3679.22 110.02 0.00 0.70 0.93 0.74Browse Refuge 3702.82 7 3716.82 147.62 0.00 0.68 0.92 0.64Browse Availability 3710.32 6 3722.32 153.12 0.00 0.68 0.93 0.74

Page 7: University of Toronto T-Space€¦ · Web viewCanadian Journal of Zoology XXXXXX Megan L. Hornseth and Robert Rempel. 201x. Seasonal resource selection of woodland caribou (Rangifer

Disturbance 3778.78 3 3784.78 215.58 0.00 0.65 0.93 0.64

Page 8: University of Toronto T-Space€¦ · Web viewCanadian Journal of Zoology XXXXXX Megan L. Hornseth and Robert Rempel. 201x. Seasonal resource selection of woodland caribou (Rangifer

DBS - Spirit Range

Hypothesis -2 *LL k AIC ∆AIC WeightAUC

Specificity Precision

Summer (N = 54)

Global Model 2439.7811 2461.78 0.00 1.00 0.82

0.97 0.79

Browse Refuge 2590.04 7 2604.04 142.26 0.00 0.75 0.96 0.58Browse Availability 2627.40 7 2641.40 177.62 0.00 0.74 0.96 0.65Competition 2652.20 6 2664.20 204.42 0.00 0.74 0.96 0.71Disturbance - 3 - - - 0.58 - -

Page 9: University of Toronto T-Space€¦ · Web viewCanadian Journal of Zoology XXXXXX Megan L. Hornseth and Robert Rempel. 201x. Seasonal resource selection of woodland caribou (Rangifer

HBL – Missisa Range

Hypothesis -2 *LL k AIC ∆AICWeigh

t AUCSpecificity Precision

Spring (N = 56)Global Model 3111.30 9 3129.30 0.00 1.00 0.73 0.99 0.96Competition 3141.70 6 3153.70 24.40 0.00 0.72 0.96 0.86Disturbance 3150.36 3 3156.36 27.06 0.00 0.71 0.95 0.78Browse Refuge 3266.92 6 3278.92 149.62 0.00 0.69 0.94 0.68Browse Availability 3300.52 5 3310.52 181.22 0.00 0.67 0.94 0.73Summer (N = 52)Global Model 4200.10 9 4218.10 0 1.00 0.73 0.96 0.92Browse Refuge 4268.70 6 4280.70 62.60 0.00 0.72 0.91 0.72Competition 4328.68 6 4340.68 122.58 0.00 0.71 0.91 0.75Browse Availability 4360.78 5 4370.78 152.68 0.00 0.71 0.88 0.59Disturbance 4548.72 3 4554.72 336.62 0.00 0.64 0.91 0.83Fall (N = 44)Global Model 3962.60 9 3980.60 0.00 1.00 0.74 0.92 0.70Competition 4036.68 6 4048.68 68.08 0.00 0.73 0.92 0.70Browse Refuge 4066.04 6 4078.04 97.44 0.00 0.74 0.92 0.71Browse Availability 4207.62 5 4217.62 237.02 0.00 0.68 0.90 0.67Disturbance 4216.94 3 4222.94 242.34 0.00 0.68 0.91 0.72Winter (N = 37)Global Model 3448.38 9 3466.38 0.00 1.00 0.83 0.95 0.78Browse Refuge 3595.90 6 3607.90 141.52 0.00 0.82 0.95 0.78Browse Availability 3619.40 5 3629.40 163.02 0.00 0.82 0.94 0.77Competition 4073.42 6 4085.42 619.04 0.00 0.73 0.92 0.74Disturbance 4384.08 3 4390.08 923.7 0.00 0.65 0.91 0.76

Page 10: University of Toronto T-Space€¦ · Web viewCanadian Journal of Zoology XXXXXX Megan L. Hornseth and Robert Rempel. 201x. Seasonal resource selection of woodland caribou (Rangifer

HBL – James Bay Range

Hypothesis -2 *LL k AIC ∆AICWeigh

t AUCSpecificity Precision

Spring (N = 26)Global Model 2720.70 9 2738.70 0.00 0.66 0.76 0.97 0.83Browse Availability 2730.94 5 2740.94 2.24 0.25 0.75 0.98 0.88Browse Refuge 2731.04 6 2743.04 4.34 0.09 0.75 0.98 0.89Competition 2783.78 6 2795.78 57.08 0.00 0.73 0.96 0.78Disturbance - 3 - - - 0.44 - -Summer (N = 28)Global Model 3362.86 9 3378.86 0.00 1.00 0.74 0.96 0.78Browse Availability 3398.38 5 3408.38 29.52 0.00 0.73 0.96 0.84Browse Refuge 3396.46 6 3408.46 29.60 0.00 0.73 0.95 0.76Competition 3421.42 6 3433.42 54.56 0.00 0.72 0.95 0.77Disturbance - 3 - - - 0.48 - -Fall (N = 26)Global Model 3932.2 9 3950.20 0.00 1.00 0.68 0.95 0.87Browse Availability 4067.42 5 4077.42 127.22 0.00 0.70 0.95 0.86Browse Refuge 4067.38 6 4079.38 129.18 0.00 0.70 0.94 0.83Competition 4098.54 6 4110.54 160.34 0.00 0.67 0.94 0.83Disturbance 4207.14 3 4213.14 262.94 0.00 0.60 0.89 0.64Winter (N = 23)Global Model 3567.14 9 3585.14 0 1.0 0.67 0.94 0.72Browse Refuge 3589.94 6 3601.94 16.80 0.00 0.66 0.93 0.75Browse Availability 3615.86 8 3631.86 21.38 0.00 0.65 0.95 0.84Competition 3679.52 6 3691.52 106.38 0.00 0.61 0.93 0.76Disturbance - 3 - - - 0.53 - -

Page 11: University of Toronto T-Space€¦ · Web viewCanadian Journal of Zoology XXXXXX Megan L. Hornseth and Robert Rempel. 201x. Seasonal resource selection of woodland caribou (Rangifer

Literature Cited

Bowman, J., Ray, J.C., Magoun, A.J., Johnson, D.S., and Dawson, F.N. 2010. Roads, logging, and the large-mammal community of an eastern Canadian boreal forest. Can. J. Zool. 88(5): 454-467.

Bradshaw, C.J.A., Hebert, D.M., Rippin, A.B., and Boutin, S. 1995. Winter peatland habitat selection by woodland caribou in northeastern Alberta. Can. J. Zool. 73(8): 1567-1574.

Dussault, C., Pinard, V., Ouellet, J.P., Courtois, R., and Fortin, D. 2012. Avoidance of roads and selection for recent cutovers by threatened caribou: fitness-rewarding or maladaptive behaviour? Proc. R. Soc. Lond. B Biol. Sci. No. 279(1746): 4481-4488.

Dyer, S.J., O'Neill, J.P., Wasel, S.M., and Boutin, S. 2002. Quantifying barrier effects of roads and seismic lines on movements of female woodland caribou in northeastern Alberta. Can. J. Zool. 80(5): 839-845.

Fortin, D., Courtois, R., Etcheverry, P., Dussault, C., and Gingras, A. 2008. Winter selection of landscapes by woodland caribou: behavioural response to geographical gradients in habitat attributes. J. Appl. Ecol. 45(5): 1392-1400.

Racey, G.D., Harris, A., Gerrish, L., Armstrong, E., Nicol, J., and Baker, J. 1999. Forest management guidelines for the conservation of woodland caribou: a landscape approach. Available from the Ontario Ministry of Natural Resources and Forestry, Thunder Bay, Ontario.

Rettie, W.J., and Messier, F. 2000. Hierarchical habitat selection by woodland caribou: its relationship to limiting factors. Ecography, 23(4): 466-478.

Vors, L.S., Schaefer, J.A., Pond, B.A., Rodgers, A.R., and Patterson, B.R. 2007. Woodland caribou extirpation and anthropogenic landscape disturbance in Ontario. J. Wildl. Manage. 71(4): 1249-1256.

Page 12: University of Toronto T-Space€¦ · Web viewCanadian Journal of Zoology XXXXXX Megan L. Hornseth and Robert Rempel. 201x. Seasonal resource selection of woodland caribou (Rangifer

Figure S1. Relative seasonal use of select environmental variables across 6 ranges by caribou

(Rangifer tarandus caribou) in northern Ontario, Canada, for (A) Mixed Forest, (B) Eskers, (C)

Sparse Treed Forest, and (D) Open Peatland. Cover types are expressed as relative proportion,

eskers are expressed as m x ha x 10-1.

Page 13: University of Toronto T-Space€¦ · Web viewCanadian Journal of Zoology XXXXXX Megan L. Hornseth and Robert Rempel. 201x. Seasonal resource selection of woodland caribou (Rangifer

Figure S2. Relative seasonal use of select environmental variables across 6 ranges by caribou

(Rangifer tarandus caribou) in northern Ontario, Canada, for (A) Water and (B) Natural

Disturbance, expressed as relative proportion.