Effects of Habitat Edges on Wintering Bird Occupancy in Umstead - - PowerPoint PPT Presentation

Effects of Habitat Edges on Wintering Bird Occupancy in Umstead State Park

AEC501 Lab Project Group 3: Connor Winfield, Jackie Hausle, Paul Lasley, Rebecca Goodnight, Sabrina Dixon, and Whitney McCurry

Study Aims

• To further explain the effects of habitat fragmentation on urbanized

landscapes.

• Analyze the differences of species composition from edge habitat (boundary)

to core habitat (middle of park).

• Expand on previous research (Murcia 1995) of edge effects on species

composition.

• Specifically focused on wintering bird populations in William B. Umstead State

Park.

Background

William B. Umstead State Park

• Located in Northwest Raleigh
• 5,600 acres of old growth forest
• 3 lakes with various creeks interspersed
• US-70 to the North
• Interstate I-40 to the South
• Raleigh-Durham International to the West
• Suburbs to the East

Question

• What are the effects on occupancy for various bird species as the distance

from the park boundary increases and the surroundings shifts to a less developed core habitat?

Methods: Collection

• Three collection days

○ Mar 17 ○ Mar 18 ○ Third rep weekday between Mar 19-Mar 23

• 3 trails - Cedar Ridge, Graylyn, and

Company Mill- consistent observers

• 22 points on each trail, 100 meters apart
• Unlimited radius
• Recorded calls or sightings,

presence/absence data

• Made note of number of visitors seen on

each trail

• Flaws?

Methods: Analysis

• Presence 12.7
• Focused on 11 species
• 3 models

○ Null: No Covariates ○ Distance: Distance from boundary into park, main study variable ○ Trail Use: Comparing Heavily traversed trail to less used trails ■ Confounded with observer and trail variable

• Statistics: based on maximum model likelihood estimates

○ AIC: Model selection metric ○ ΔAIC: Measurement from the top-ranked model’s AIC ○ Weight: Measure of support for each model being the best model

Results

Overview

• 39 species found

○ 34 Passerines ○ 4 Piciformes ○ 2 Accipitriformes

• 11 chosen for analysis

Naive Occupancy Compared to Expected Occupancy

• Presence adjusts for

detection

○ Based on percentage of replicates where species was detected

• Disparity between

Naive/Calculated indicates detection probability

○ Inverse relation

Model Reliability

American Robin Mourning Dove AIC

ΔAIC

Weight AIC

ΔAIC

Weight Null 147.03 0.857 Null 181.26 0.463 Distance 152.36 5.33 0.0596 Distance 182.73 1.47 0.222 Trail Use 151.69 4.66 0.0834 Trail Use 182.03 0.77 0.315 Brown-headed Nuthatch Northern Cardinal AIC

ΔAIC

Weight AIC

ΔAIC

Weight Null 117.27 3.3 0.1204 Null 218.73 6.67 0.0356 Distance 113.97 0.6271 Distance 220.19 8.13 0.0172 Trail Use 115.79 1.82 0.2524 Trail Use 212.06 0.9499

Model Reliability

Carolina Wren Pine Warbler AIC

ΔAIC

Weight AIC

ΔAIC

Weight Null 197.64 16.35 0.0003 Null 222.44 10.47 0.0053 Distance 199.48 18.19 0.0001 Distance 238.68 26.71 Trail Use 181.29 0.9996 Trail Use 211.97 0.9947 Dark Eyed Junco Red-Bellied Woodpecker AIC

ΔAIC

Weight AIC

ΔAIC

Weight Null 128.36 24.12 Null 153.63 1.4 0.2956 Distance 126.13 21.89 Distance 155.62 3.39 0.1093 Trail Use 104.24 1 Trail Use 152.23 0.5952

Model Reliability

Eastern Towhee Tufted Titmouse AIC

ΔAIC

Weight AIC

ΔAIC

Weight Null 127 27.47 Null 182.26 4.57 0.0824 Distance 128.57 29.04 Distance 191.71 4.02 0.1084 Trail Use 99.53 1 Trail Use 177.69 0.8092 Golden Crowned Kinglet AIC

ΔAIC

Weight Null 155.39 45.04 Distance 157.02 46.67 Trail Use 110.35 1

Model Reliability: Discussion

• Preferred Model has lowest AIC

○ Preferred compared to other models, not to best possible model

• Our Results:

○ Trail Use: 8 ○ Distance: 1 ○ Null: 2

Occupancy v. Distance

• Biological significance

○ Edge species ○ Interior species

• Missing species due to

presence parameters being violated

Species Breakdown Occupancy v. Distance

Species Breakdown Occupancy v. Distance

Species Breakdown Occupancy v. Distance

Discussion

• Trail use model preferred in 8 species

○ 2 light trails, 1 heavy trail: not enough replicates for significant analysis ○ Confounded with observer skill variable as well

• Surprising detection probability results?

○ Robin detection much lower than Dark Eyed Junco? ■ Why?

• Observer Bias
• Sampling Bias

Discussion

• Estimates of occupancy at

set distances

○ Some species had naive

• ccupancy of 1 (across 9 reps

for each set distance) ○ So can’t run analysis

• Most statistically insignificant

○ American Robin, Dark-Eyed Junco show statistically significant differences in

• ccupancy between outermost

and innermost data points

Acknowledgements

• Ted Simons for 99% of our avian identification knowledge
• Shilo Felton and Mike Cove for Presence Instruction and Troubleshooting
• Shilo Felton for constructive criticism along the way

Literature Cited

Blair, R. B. (1996), Land Use and Avian Species Diversity Along an Urban Gradient. Ecological Applications, 6: 506–

• 519. doi:10.2307/2269387

Ewers, R. and Didham, R. (2005). Confounding factors in the detection of species responses to habitat

• fragmentation. Biological Reviews, 81(01), p.117.

Murcia, C.1995. Edge Effects in Fragmented Forests: Implications for Conservation. Trends in Ecology & Evolution 10:58- 62. Zurita, G., Pe’er, G., Bellocq, M. and Hansbauer, M. (2012). Edge effects and their influence on habitat suitability calculations: a continuous approach applied to birds of the Atlantic forest. Journal of Applied Ecology, 49(2), pp.503-512.