proximity to unconventional shale gas infrastructure

AmericanOrnithology.org

Volume 121, 2019, pp. 1–20DOI: 10.1093/condor/duz020

Copyright © American Ornithological Society 2019. All rights reserved. For permissions, e-mail: [email protected].

RESEARCH ARTICLE

Proximity to unconventional shale gas infrastructure alters breeding bird abundance and distributionLaura S. Farwell,1,a,* Petra B. Wood,2 Donald J. Brown,1,3 and James Sheehan1

1 West Virginia Cooperative Fish and Wildlife Research Unit, West Virginia University, Morgantown, West Virginia, USA2 U.S. Geological Survey, West Virginia Cooperative Fish and Wildlife Research Unit, West Virginia University, Morgantown, West Virginia, USA3 U.S. Forest Service, Northern Research Station, Parsons, West Virginia, USAa Current Address: Department of Forest and Wildlife Ecology, University of Wisconsin-Madison, Madison, Wisconsin, USA* Corresponding author: [email protected]

Submission Date: October 31, 2018; Editorial Acceptance Date: April 3, 2019; Published 11 June 2019

ABSTRACTUnconventional shale gas development is a rapidly expanding driver of forest loss and fragmentation in the cen-tral Appalachian region. We evaluated the relationship between breeding passerine abundances and distance from shale gas development at a long-term (2008–2017) study site in northern West Virginia, USA. We examined responses of 27 species within 3 habitat guilds: forest interior, early successional, and synanthropic. More than half of the species evaluated showed sensitivity to distance from unconventional shale gas infrastructure (e.g., well pads, ac-cess roads, pipelines). Five forest interior species occurred in greater abundances farther from shale gas develop-ment, whereas 3 forest interior gap specialists increased in abundance closer to shale gas. Early successional and synanthropic species, including the nest-parasitic Brown-headed Cowbird (Molothrus ater), generally occurred in greater abundances closer to shale gas infrastructure. We used interpolated distributions of 4 focal species to assess their spatial response to unconventional shale gas development over time. Our results indicate that breeding pas-serine distributions and community composition are changing with forest disturbance driven by unconventional shale gas energy development.

Keywords: Appalachians, avian guilds, energy development, forest songbirds, hydraulic fracturing, land-use change, Marcellus-Utica, unconventional shale gas

La proximidad a infraestructura de gas de esquisto no convencional altera la abundancia y distribución de las aves reproductivas

RESUMENEl desarrollo de gas de esquisto es un impulsor en rápida expansión de la pérdida y fragmentación del bosque en la región de los Apalaches centrales. Evaluamos la relación entre la abundancia de paserinos reproductivos y la distancia al desarrollo de gas de esquisto en un sitio de estudio de largo plazo (2008–2017) en el norte de West Virginia, EEUU. Examinamos las respuestas de 27 especies dentro de tres gremios de hábitat: interior de bosque, sucesión temprana y sinantrópico. Más de la mitad de las especies evaluadas mostraron sensibilidad a la distancia desde la infraestructura de gas de esquisto no convencional (e.g., plataformas de pozos, caminos de acceso, gasoductos). Cinco especies de interior de bosque estuvieron presentes en mayores abundancias a más distancia del desarrollo del gas de esquisto, mientras que tres especialistas de claros de interior de bosque aumentaron su abundancia más cerca del gas de esquisto. Las especies de la sucesión temprana y las sinantrópicas, incluyendo la especie parasitaria de nido Molothrus ater, generalmente se presentaron a mayores abundancias más cerca de la infraestructura de gas de esquisto. Usamos distribuciones interpoladas de cuatro especies focales para evaluar su respuesta espacial al desarrollo de gas de esquisto no convencional a lo largo del tiempo. Nuestros resultados indican que las distribuciones y la composición de la comunidad de los paserinos reproductivos están cambiando con el disturbio del bosque generado por el desarrollo del gas de esquisto no convencional.

Palabas clave: Apalaches, aves canoras del bosque, cambio de uso del suelo, desarrollo de energía, fractura hidráulica, gas de esquisto no convencional, gremios de aves, Marcellus-Utica

mailto:[email protected]?subject=

2 Passerine response to distance from fracking infrastructure L. S. Farwell, et al.

The Condor: Ornithological Applications 121:1–20, © 2019 American Ornithological Society

INTRODUCTION

Human conversion of land cover and resulting loss and fragmentation of habitat are major drivers of global de-clines in bird populations (Gaston et al. 2003, Jetz et al. 2007). Increasing global energy demands and techno-logical advances in fossil fuel extraction have resulted in a rapid expansion of energy development in North America in recent decades (USEIA 2015, 2018). Energy sector dis-turbance often occupies a relatively small proportion of land cover, but creates a disproportionate amount of edge habitat from a proliferation of linear infrastructure (Bayne and Dale 2011). Increased edge creation from energy de-velopment degrades habitat suitability for some edge-avoiding species, as evidenced by declines in area-sensitive birds in response to seismic lines in boreal forests (Bayne et al. 2005, Machtans 2006), and oil and gas development in sagebrush steppe (Gilbert and Chalfoun 2011, Mutter et al. 2015) and grasslands (Thompson et al. 2015, Nenninger and Koper 2018).

The central Appalachian region in the eastern United States has seen a boom in drilling for shale gas since the mid-2000s, owing in part to advances in horizontal drilling and high-volume hydraulic fracturing technologies (i.e. fracking; USEIA 2018). This method of extracting natural gas from deep shale formations is relatively new to the eastern United States, but there is growing evidence that forest loss and fragmentation associated with unconventional shale gas development is negatively affecting interior forests in the region (Drohan et al. 2012, Langlois et al. 2017) and re-structuring avian communities (Barton et al. 2016, Farwell et al. 2016, Langlois 2017). Understanding how birds react spatially to unconventional shale gas infrastructure within forested environments is a key step in predicting how avian populations will respond to large-scale land use change driven by this emerging form of energy development.

Much of the unconventional shale gas development (hereafter, “shale gas”) in the central Appalachian region is occurring in interior forests where pipelines and access roads result in extensive forest fragmentation (Drohan et al. 2012, Farwell et al. 2016, Langlois et al. 2017). Shale gas is negatively affecting abundance and diversity of area-sensitive forest interior birds (Barton et al. 2016, Farwell et al. 2016, Langlois 2017), as well as the reproductive suc-cess of a forest interior, riparian obligate species (Frantz et al. 2018b). At the same time, some early successional and synanthropic species appear to benefit from forest disturbance by shale gas (Barton et al. 2016, Farwell et al. 2016, Langlois 2017). Further complicating the issue, edge-associated species and forest gap specialists are naturally drawn to canopy openings, and small-scale anthropogenic disturbances can mimic natural canopy openings that then function as ecological traps (Weldon and Haddad 2005, Boves et al. 2013b).

We assessed the relationship between passerine abun-dances and distance from shale gas at a long-term study site in northern West Virginia. We also created annual in-terpolated distribution maps for 4 species to estimate the spatial relationship between bird abundances and locations of shale gas infrastructure. Our objective was to determine if proximity to shale gas influenced patterns in bird abun-dance and local distribution. Our results will help quantify the spatial effects of shale gas on birds and inform spatially explicit management objectives for species in landscapes undergoing shale gas development.

METHODS

Study AreaLewis Wetzel Wildlife Management Area (LWWMA) in northern West Virginia overlies the Marcellus and Utica shale gas basins (Figure 1). The site is located in the Permian Hills subdivision of the Western Allegheny Plateau ecoregion and is representative of the surrounding region, characterized by steep, dendritic topography and mature Appalachian oak and mixed mesophytic forests. Approximately two-thirds of the LWWMA (4,326 ha) was selected for study, with surveys located on most ridges and 50.4 km of first- and second-order tributary streams (Figure 1). Unconventional shale gas development began at the site in 2007 and expanded steadily from 2008 to 2015 (Figure 2). Although activity slowed at the site with no new wells drilled during 2014–2017, existing wells continue to produce gas at the site, and a 25-ha permitted drilling area and access road were cleared in 2015. As of 2017, the study area remained predominantly (90.0%) forested, but it has been heavily fragmented by the construction of gas well pads and, to a greater extent, by the proliferation of new road and pipeline corridors (Farwell et al. 2016).

Land Cover Mapping and Distance MetricsWe manually digitized and classified all land cover at the site as vector polygons from 2008 to 2017, from a sequence of leaf-off and leaf-on aerial and satellite imagery (ArcGIS 10.4; ESRI 2016; see Appendix Table 3 for list of image sources). Land cover classes included forest, timber har-vest, shale gas, and other forms of human development un-related to shale gas. Preexisting impacts (i.e. roads, utility corridors, and private inholdings) were delineated and treated as baseline disturbance because they were present on the landscape 20+ years prior to the start of the study (Figure 1). Three partial timber harvests (24 ha total) that occurred prior to the study (2006–2007) were classified as background forest since most canopy trees were retained, and the harvests resulted in minimal impacts to the forest bird community at this study site (Sheehan et al. 2013). Changes in land cover were digitized annually and cat-egorized by type of disturbance. For disturbances related

L. S. Farwell, et al. Passerine response to distance from fracking infrastructure 3


to shale gas, we grouped impacts into those associated with drilling sites or “well pads” (i.e. pads, buffers, fluid impoundments, storage areas) vs. linear forms of infra-structure (i.e. pipelines, access roads). We then calculated distance (m) from each avian sampling point to nearest shale gas polygon edge with the NEAR function (ArcGIS 10.4; ESRI 2016).

For our analyses, we focused on distance to shale gas be-cause it was the dominant source of land cover change (175

ha total) at the site during the study period. A shelterwood harvest (7 ha) and light selection harvest (59 ha) with most canopy trees retained occurred prior to the 2011 breeding season, and an even-aged harvest (17 ha) occurred prior to the 2016 breeding season. However, there are several indications that shale gas remained the primary driver of changes we observed in avian abundance and distribution from 2008 to 2017. First, 61 of 142 survey points were af-fected by shale gas within 100 m, compared with only 7

FIGURE 1. Manually digitized vector polygons representing preexisting human impacts (20+ years old) and all shale gas development at the study area as of 2017. Dots represent bird survey locations on ridges (black) and streams (blue). The black X at the south end of the site represents the point we were unable to survey in 2014. (Inset map) White star represents the study site location; shaded red area shows the extent of the Marcellus shale formation in the eastern United States.



points affected by harvests within 100 m, and 5 of these were closer to shale gas (i.e. nearest edge of human dis-turbance was shale gas, not harvest). Second, forest in-terior birds showed declines prior to 2011 when the first harvests occurred and only showed significant declines at points impacted by shale gas within 100 m (Farwell et al. 2016). Third, regenerating harvests form a dynamic mo-saic with surrounding forests and generally have fewer negative effects on forest interior species compared with more permanent, constructed disturbances (Schmiegelow et al. 1997, Hansen and Rotella 2000, Lichstein et al. 2002). Harvests at our study site formed less abrupt land cover transitions than shale gas, with retention and regeneration of native vegetation and many residual trees, post-harvest.

Avian Surveys and Detection ProbabilityDuring 2008–2017, we conducted standard, 10-min avian point count surveys (Bibby et al. 1992) at 142 survey sta-tions. We placed survey locations systematically in a GIS at ≥250-m intervals to avoid counting individual birds at more than 1 sampling point (Figure 1). We surveyed a fairly balanced sample of points located on ridges (n = 87) and along streams (n = 55) to control for variation in species abundances driven by topographic position. Each point was typically surveyed twice per breeding season. However, we missed second within-season surveys at 4 points (each in different years), and were unable to survey 1 point in 2014 owing to noise interference from shale gas activity. We conducted a total of 2,834 surveys during 2008–2017 and used maximum counts from within-year visits (n = 1,419) to minimize observer effects and false zeros (Ralph et al. 1995). We also restricted our data to observations within a 50-m radius (0.79 ha) of point count locations to fur-ther reduce observer effects and potential inconsistencies

in distance estimation (Alldredge et al. 2007). Over the course of the study, 35 observers conducted surveys; 13 conducted surveys in multiple years and 5 observers com-pleted 60% of all counts. All observers were proficient in regional bird identification and distance estimation, and were trained and field-tested prior to surveying.

We conducted surveys between May 18 and July 5, from sunrise until 4 hr after sunrise on mornings with suitable weather and environmental conditions. First detections of birds were recorded within 7 time intervals (0–2, >2–3, >3–4, >4–5, >5–6, >6–7, >7–10 min) and 2 distance bands (0–25, >25–50 m) for estimation of detection probability with a combined removal and distance sampling method (Sólymos et al. 2013). This method provides conditional likelihood estimates for availability and perceptibility, the 2 components of detectability (Sólymos et al. 2013). We in-cluded serial date, time since sunrise, and quadratic terms of each as survey-level covariates potentially influencing availability (probability that a bird will sing). We included observer experience level (a ranking of 1–5, based on rela-tive number of years of experience conducting avian point count surveys) and tree cover surrounding each point as survey-level covariates potentially influencing percepti-bility (probability that a bird will be detected if it sings). Tree cover was estimated as percent forest cover within a 100-m radius of each survey point. We converted annual vector land cover maps (described above) to 1-m reso-lution raster grids and used Fragstats 4 (McGarigal et al. 2012) to calculate percent forest cover within 100-m buf-fers of survey locations.

We grouped detected species into 3 avian habitat guilds likely to be affected by human development in the Appalachian region (Thomas et al. 2014, Farwell et al. 2016). Forest interior birds are associated with large areas of mature forest; early successional birds prefer re-cently disturbed, shrubby habitat; synanthropic birds are habitat generalists known to thrive in human-modified environments.

Data Analysis and InterpolationWe fitted 9 removal models and 4 distance sampling models for each species, including intercept-only models (Appendix Table 4) with the package DETECT (Sólymos et al. 2016) in program R 3.4.1 (R Core Team 2017). Fitted models with smallest AIC values were selected and used to calculate species-specific correction factors to account for detection probability, and applied as custom offsets in the R program package lme4 (Sólymos et al. 2013, Bates et al. 2015). We related species abundances within 50-m radius buffers (0.79 ha) of survey points to distance from shale gas development with generalized linear mixed models (GLMM) in lme4, with a Poisson distribution and log link (Bates et al. 2015). We standardized predictor

FIGURE 2. Site-wide increases in shale gas footprint over time, for all shale gas combined and separately for linear infrastructure (i.e. pipelines, access roads) vs. well pad areas (i.e. pads, buffers, fluid impoundments, storage areas). Dotted line represents linear trend obtained from ordinary least squares regression.



variables prior to modeling, as suggested by Bates et al. (2015), to improve model convergence. During explora-tory data analysis, we checked for overdispersion and zero-inflation of abundance data (Zuur et al. 2010). Blue Jay (see Table 1 for scientific names) was the only species with slight zero-inflation; a comparison of log-likelihood ratios revealed negative binomial models were a better fit than Poisson models, so for Blue Jay we applied a nega-tive binomial distribution in lme4 (Bates et al. 2015). For all species, we originally modeled both linear and quad-ratic terms of distance from shale gas to assess potential nonlinear responses. However, we found little evidence of quadratic responses among species evaluated, and found stronger support for linear responses in the few species that also showed a quadratic response. Thus, we only in-cluded linear terms in final models.

First, we modeled bird abundances at all points relative to nearest distance from any shale gas (n = 1,419), then separately for only those points nearest to pads (n = 634) vs. those nearest to linear infrastructure (n = 785). We in-cluded point and year as random effects in all GLMMs; models contained random intercepts but not random slopes because we were most interested in controlling for variation associated with repeated measures over space and time, not in variation in bird response among points and years (Zuur et al. 2009). To evaluate GLMM goodness of fit, we calculated conditional R2 values, which measure the proportion of variance explained by both fixed and random effects (Nakagawa et al. 2017), in the R package SJSTATS (Lüdecke 2019).

We considered using a before–after control–impact (BACI) design with this dataset. However, there was a prob-lematic lack of clear distinction between “before” and “after” due to many points being impacted multiple times over the long-term study period (Smith et al. 1993). Designating in-dividual points as “control” vs. “impact” was not straightfor-ward given the unknown spatial extent of shale gas impacts on birds. Since this was a central question of the current study, we chose instead to focus on a point-level assessment of effects of distance from gas on avian abundance.

To evaluate differing responses among species to dis-tance from shale gas, we plotted predicted abundances for 4 species based on deterministic models (population-level, fixed effects only) using the PREDICT function (R Core Team 2017). We plotted predicted responses for 2 species that decreased in abundance (Ovenbird, Wood Thrush) and 2 species that increased in abundance (Brown-headed Cowbird, Indigo Bunting) closer to shale gas. We then computed 95% confidence intervals for predictive models using 1,000 bootstrap replications.

Last, we compared distributions of 4 focal species across the study area to locations of shale gas with spatially con-tinuous distribution probability maps, interpolated with

kriging (ArcGIS 10.4; ESRI 2016). We selected 4 species representing a range of habitat preferences and resource needs (Ehrlich et al. 1988, Rodewald 2015). The Ovenbird is an area-sensitive species that breeds in extensive, closed-canopy forests with open understory and deep leaf litter. The Cerulean Warbler is also associated with large areas of mature forest but prefers small canopy gaps and open-ings within extensive forests (Wood et al. 2006, Boves et al. 2013b). The Indigo Bunting is an early successional spe-cies typically found in shrubby, weedy habitats between forests and open areas. The Brown-headed Cowbird is a synanthropic, nest-parasitic species that is particularly at-tracted to human disturbance and linear corridors (Gates and Evans 1998, Howell et al. 2007) and has been impli-cated in declines of multiple species of conservation con-cern (Eckrich et al. 1999, Kus 1999, DeCapita 2000).

Kriging is a family of regionalized variable interpolation techniques based on regression against observed values of surrounding data points, weighted according to spatial covariance values (Li and Heap 2008). Indicator kriging (IK) is a nonparametric approach that transforms con-tinuous data to binary values that indicate membership to classes, based on observed values within a defined neigh-borhood (Li and Heap 2008). Although IK often is used in ecological studies because it does not require normally distributed data, a disadvantage of IK is that it cannot be used to generate predictive maps of species abundance (Walker et al. 2008). Rather, IK maps species distributions based on probability of a species’ occurrence. Probability of occurrence maps are not suitable for accurate estima-tion of animal densities, but can be used as indices of rela-tive abundance to evaluate patterns of species distributions (Walker et al. 2008, Melles et al. 2011, Polakowska et al. 2012). During exploratory analysis of abundance data from 2008 to 2017, we compared performance of IK with 2 other interpolation methods: ordinary kriging and inverse dis-tance weighting (Li and Heap 2008, Abdi and Nandipati 2009). We found that IK consistently outperformed the other 2 interpolation methods evaluated, and thus used IK to interpolate distribution probability maps for focal spe-cies (see Appendix Table 5 for cross-validation results from the 3 interpolation methods). We used root mean square of standardized error (RMSSE) to evaluate the performance of IK interpolations for the 4 focal species (Polakowska et al. 2012). This measure of interpolation performance in-dicates how well estimated and measured variances match, and optimally should be 1 (Johnston et al. 2003).

RESULTS

Avian Response to Distance from Gas DevelopmentOver the 10-yr study period, shale gas steadily expanded in the study area and encroached on interior forest



TAB

LE 1

. G

ener

aliz

ed li

near

mix

ed m

odel

out

puts

show

ing

rela

tions

hips

bet

wee

n sp

ecie

s abu

ndan

ce (n

umbe

r per

0.7

9 ha

) and

dis

tanc

e to

shal

e ga

s (m

), gr

oupe

d by

hab

itat

guild

s: fo

rest

inte

rior (

FI),

early

suc

cess

iona

l (ES

), an

d sy

nant

hrop

ic (S

YN).

Resu

lts a

re s

how

n fo

r all

poin

ts re

lativ

e to

nea

rest

sha

le g

as o

f any

type

(“A

ll sh

ale

gas

com

bine

d”;

n =

1,41

9), t

hen

sepa

rate

ly fo

r poi

nts

near

est t

o w

ell p

ads

(n =

634

) vs.

linea

r inf

rast

ruct

ure

(n =

785

). Po

sitiv

e es

timat

es in

dica

te g

reat

er a

bund

ance

s w

ith in

crea

sing

dis

tanc

e fr

om s

hale

gas

. Sig

nific

ant r

espo

nses

are

in b

old

(α =

0.0

5), a

nd c

ondi

tiona

l R2 v

alue

s re

pres

ent v

aria

nce

expl

aine

d by

bot

h fix

ed a

nd ra

ndom

effe

cts

(Nak

agaw

a et

al.

2017

).

All

shal

e ga

s co

mbi

ned

Gas

wel

l pad

sLi

near

gas

infr

astr

uctu

re

Gui

ld S

peci

esEs

t.SE

zP

R2Es

t.SE

zP

R2Es

t.SE

zP

R2

FIEa

ster

n W

ood-

Pew

ee a

Cont

opus

vire

ns−0

.27

0.09

−2.9

40.

003

0.32

−0.3

50.

15−2

.27

0.02

0.35

−0.2

20.

10−2

.11

0.03

0.30

Ac

adia

n Fl

ycat

cher

a Em

pido

nax

vire

scen

s0.

010.

070.

190.

85−

0.00

40.

140.

030.

97−

0.05

0.08

0.66

0.51

−

Bl

ue-h

eade

d Vi

reo

Vi

reo

solit

arie

s0.

040.

090.

380.

70−

−0.1

10.

17−0

.69

0.49

−−0

.03

0.11

−0.2

30.

82−

Re

d-ey

ed V

ireo

Vi

reo

oliv

aceu

s0.

070.

032.

610.

010.

100.

100.

042.

610.

010.

120.

100.

032.

780.

010.

13

W

ood

Thru

sh a

,b,c

Hyl

ocic

hla

mus

telin

a0.

080.

051.

570.

12−

−0.0

30.

07−0

.53

0.60

−0.

180.

062.

940.

003

0.33

O

venb

ird

Seiu

rus a

uroc

apill

a0.

250.

064.

11<0

.001

0.33

0.18

0.08

2.23

0.03

0.26

0.30

0.08

3.56

<0.0

010.

36

W

orm

-eat

ing

War

bler

a,b

H

elm

ither

os v

erm

ivor

um0.

050.

090.

550.

58−

−0.0

40.

11−0

.37

0.71

−0.

120.

111.

150.

25−

Lo

uisi

ana

Wat

erth

rush

a

Park

esia

mot

acill

a−0

.11

0.11

−0.9

80.

33−

−0.3

30.

18−1

.84

0.07

−−0

.03

0.14

−0.2

00.

85−

Bl

ack-

and-

whi

te W

arbl

er a

Seto

phag

a ca

erul

esce

ns−0

.05

0.05

−1.0

20.

31−

0.06

0.06

1.07

0.28

−−0

.12

0.06

−1.8

40.

07−

Ke

ntuc

ky W

arbl

er a

,b

Geo

thly

pis f

orm

osa

−0.0

70.

10−0

.72

0.47

−−0

.17

0.17

−0.9

90.

32−

0.07

0.11

0.64

0.52

−

H

oode

d W

arbl

er a

Seto

phag

a ci

trin

a0.

050.

041.

330.

18−

0.12

0.05

2.48

0.01

0.12

−0.0

30.

05−0

.63

0.53

−

A

mer

ican

Red

star

t Se

toph

aga

rutic

illa

−0.0

90.

05−1

.71

0.09

−−0

.23

0.08

−2.7

00.

010.

34−0

.03

0.06

−0.4

10.

68−

Ce

rule

an W

arbl

er a

,b,c

Seto

phag

a ce

rule

a−0

.19

0.06

−3.0

10.

003

0.35

−0.0

50.

09−0

.50

0.62

−−0

.20

0.08

−2.6

30.

010.

35

Bl

ack-

thro

ated

Gre

en

War

bler

Se

toph

aga

vire

ns

0.20

0.10

2.00

0.05

0.16

0.12

0.18

0.68

0.50

−0.

140.

101.

340.

18−

Sc

arle

t Tan

ager

a

Pira

nga

oliv

acea

0.02

0.04

0.54

0.59

−0.

100.

051.

880.

06−

0.00

30.

050.

060.

95−

Ro

se-b

reas

ted

Gro

sbea

k Ph

euct

icus

ludo

vici

anus

0.09

0.07

1.25

0.21

−0.

160.

101.

630.

10−

0.03

0.08

0.32

0.75

−

ESG

ray

Catb

ird

Dum

etel

la ca

rolin

ensi

s−0

.22

0.35

−0.6

20.

54−

−0.2

20.

35−0

.62

0.54

−0.

030.

490.

060.

95−

Ea

ster

n To

whe

e a

Pipi

lo e

ryth

roph

thal

mus

−0.0

20.

05−0

.36

0.72

−0.

010.

080.

190.

85−

−0.0

50.

05−0

.95

0.34

−

Fi

eld

Spar

row

a

Spiz

ella

pus

illa

−0.1

20.

17−0

.69

0.49

−−0

.17

0.15

−1.1

70.

24−

−0.3

00.

20−1

.48

0.14

−

Bl

ue-w

inge

d W

arbl

er a

,b

Verm

ivor

a cy

anop

tera

−0.1

00.

13−0

.76

0.45

−−0

.15

0.21

−0.7

20.

47−

−0.2

30.

19−1

.22

0.22

−



habitat (Figure 2). The total footprint of shale gas in-creased tenfold, from 17 ha in 2008 to 175 ha in 2017, with the larger proportion of this change driven by linear infrastructure (105 ha) compared with well pads (70 ha). In 2008, the mean distance from avian sampling points to the closest shale gas was 901.8 m (range: 0–2,645.5 m), and only 12% of points fell within 100 m of shale gas. Ten years later, in 2017, the mean distance to closest shale gas dropped to 333.7 m (range: 0–2,026.6 m) and 43% of points fell within 100 m of shale gas. Well pad areas averaged 1.2 ha (±0.3 SE), but some were considerably larger and ranged up to 14 ha in total area. Shale gas pipelines within the study area were generally 15–35 m in width, whereas access road corridors varied from 10 to 150 m in width, owing to complex terrain and highly variable roadside buffer widths. Thus, the distances we measured to nearest shale gas edge represented impacts of variable sizes.

We detected a total of 79 species of passerines and near-passerines within 50 m of our sampling points, excluding nonbreeding migrants and species only observed once during the 10-yr study. Based on model convergence, and as demonstrated in Sólymos et al. (2013), we found that ~75 detections were required to successfully run detection probability models (Supplemental Material Table S1) and subsequent mixed effects models. Of 36 species with ≥75 detections, 27 were passerines that fell within our 3 avian guilds of interest: forest interior (n = 16), early successional (n = 6), and synanthropic (n = 5). In our analyses we fo-cused on these 27 species, many of which are species of conservation concern (Table 1).

Of the 27 species evaluated, we identified 14 whose abun-dances were significantly related to distance from shale gas, with proportion of variance explained ranging from 0.10 to 0.59 (Table 1). Within the forest interior guild, 8 of 16 species evaluated showed a significant response. Black-throated Green Warblers, Ovenbirds (Figure 3A), and Red-eyed Vireos occurred in greater numbers with increasing distance from all types of shale gas infrastructure, whereas Hooded Warblers increased farther from well pads and Wood Thrushes (Figure 3B) increased farther from linear infrastructure. Conversely, 3 forest interior species showed a negative response to distance from gas, indicating they occurred in greater numbers closer to shale gas. Eastern Wood-Pewees increased in abundance near all types of shale gas, while American Redstarts were more abundant near well pads and Cerulean Warblers were more abun-dant near linear infrastructure.

Within the early successional guild, only 2 of 6 species showed a significant response to distance from shale gas (Table 1). Common Yellowthroats occurred in greater abundance near all shale gas, and Indigo Buntings (Figure 3D) increased in abundance closer to linear shale gas infra-structure, in particular.

All

shal

e ga

s co

mbi

ned

Gas

wel

l pad

sLi

near

gas

infr

astr

uctu

re

Gui

ld S

peci

esEs

t.SE

zP

R2Es

t.SE

zP

R2Es

t.SE

zP

R2

Co

mm

on Y

ello

wth

roat

G

eoth

lypi

s tric

has

−0.5

40.

12−4

.45

<0.0

010.

40−0

.71

0.24

−2.9

80.

003

0.27

−0.4

90.

16−3

.05

0.00

20.

36

In

digo

Bun

ting

a

Pass

erin

a cy

anea

−0.1

90.

08−2

.52

0.01

0.57

−0.1

60.

14−1

.11

0.27

−−0

.25

0.10

−2.4

80.

010.

59

SYN

Blue

Jay

Cyan

ocitt

a cr

ista

ta0.

180.

072.

510.

010.

210.

170.

062.

600.

010.

300.

140.

062.

160.

030.

29

A

mer

ican

Rob

in

Turd

us m

igra

toriu

s−0

.26

0.08

−3.2

50.

001

0.16

−0.2

60.

13−2

.03

0.04

0.21

−0.2

10.

10−2

.22

0.03

0.15

Ch

ippi

ng S

parr

ow

Spiz

ella

pas

serin

a−0

.92

0.17

−5.5

3<0

.001

0.48

−0.8

90.

33−2

.67

0.01

0.58

−0.7

00.

17−4

.09

<0.0

010.

20

Br

own-

head

ed C

owbi

rd

Mol

othr

us a

ter

−0.8

50.

17−4

.94

<0.0

010.

50−1

.12

0.16

−7.1

6<0

.001

0.51

−0.9

40.

20−4

.79

<0.0

010.

44

N

orth

ern

Card

inal

Ca

rdin

alis

card

inal

is−0

.01

0.08

−0.1

70.

87−

0.03

0.11

0.29

0.77

−−0

.11

0.10

−1.0

30.

31−

a App

alac

hian

Mou

ntai

ns Jo

int V

entu

re C

onse

rvat

ion

Prio

rity

Spec

ies

(AM

JV 2

015)

. b U

SFW

S Bi

rds

of C

onse

rvat

ion

Conc

ern,

Nat

iona

l Lis

t (U

SFW

S 20

08).

c IUCN

Red

Lis

t Spe

cies

(IU

CN 2

015)

.

TAB

LE 1

. Co

ntin

ued

https://academic.oup.com/condor/article-lookup/doi/10.1093/condor/duz020#supplementary-data



Within the synanthropic guild, 4 of 5 human-adapted species we evaluated showed significant responses to dis-tance from shale gas. Three species occurred in greater abundances closer to all shale gas (Table 1), including Brown-headed Cowbird (Figure 3C). Only Blue Jays in-creased in abundance with increasing distance from all shale gas.

Species Distribution Probability MapsThroughout the study, Ovenbirds occurred in greatest concentrations in the east-central region of the study area (Figure 4), in the largest area of interior forest undisturbed by any human development, including preexisting non-gas

development (Figure 1). However, Ovenbirds were still dis-tributed across most of the site for the first several years of the study. Beginning in 2011, we saw steadily increasing portions of the study area with a 0–0.1 probability of Ovenbird occurrence, which coincided with the largest in-creases in shale gas development during the study period (2011–2015; Figure 2). By 2014, there was a noticeable shift in Ovenbirds away from the northern area most heavily impacted by well pad development, and to a lesser extent, away from linear infrastructure throughout the study area. This spatial pattern of avoidance is supported by a 35% de-cline in Ovenbird abundance at the study site (Figure 5A), and by our distance-based models (Table 1), which showed

FIGURE 3. Examples of 2 species that increased in abundance with increasing distance (m) from unconventional shale gas devel-opment (A, B), and 2 species that decreased in abundance with increasing distance from gas development (C, D). Ovenbirds and Brown-headed Cowbirds showed significant responses to distance from all types of unconventional shale gas development (A, C), while Wood Thrush and Indigo Buntings responded to distance from linear shale gas infrastructure, in particular (B, D). Plots show model-predicted abundance, with 95% confidence intervals shaded gray; confidence intervals were computed using 1,000 bootstrap replications. Predictions are based on deterministic models at the population level and do not account for random effect variances. Predicted abundances represent abundances within a 50-m radius of sampling points (0.79 ha).



greater numbers of Ovenbirds with increasing distance from both shale gas well pads and linear infrastructure.

In contrast, Cerulean Warblers showed a less distinct pattern of spatial response to shale gas locations over time (Figure 6). Like Ovenbirds, Cerulean Warblers ini-tially occurred in some areas of forest least disturbed by any human development at the site. Unlike Ovenbirds, Cerulean Warblers also appeared more likely to occur in concentrated abundances near shale gas. Starting in 2009, Cerulean Warblers began to occur in greater abundances near pipelines and access roads, in particular. Again, this is

consistent with our distance-based models (Table 1), which showed increasing abundances of Cerulean Warblers closer to linear infrastructure, but not well pads. However, long-term abundance trends indicate Cerulean Warblers still declined across the site by 34% (Figure 5B).

Conversely, at the start of the study in 2008, Indigo Buntings occurred in high concentrations in the northern region of the study area (Figure 7), which contained the most preexisting non-gas development (Figure 1). Starting in 2009 Indigo Buntings expanded to the southern por-tion of the study site, following new shale gas construction.

FIGURE 4. Annual interpolated distribution probability maps for Ovenbirds, based on indicator kriging. GLMMs showed Ovenbirds occurred in greater abundance with increasing distance from both shale gas well pads and linear infrastructure (Table 1). In interpolated maps, darker red areas represent locations with greater probability of occurrence, and all shale gas is shown in black, for reference.



Before much of the linear shale gas infrastructure was de-veloped, Indigo Buntings were distributed across most of the site. However, as linear infrastructure expanded across the site, they gravitated toward it. Throughout the last half of the study, Indigo Buntings clearly occurred in greatest concentrations around pipelines and access roads. This

spatial pattern is supported by distance-based models (Table 1), which showed greater abundances of Indigo Buntings closer to linear shale gas infrastructure. However, long-term annual abundance trends indicate only mod-estly increasing overall numbers of Indigo Buntings across the site (Figure 5D).

FIGURE 5. Site-wide annual abundance trends (mean ±SE) for 4 focal species: 2 forest interior–dependent species that declined across the site over the study period (A, B); 1 human-adapted nest-parasitic species that increased over the same time period (C); and 1 early successional species that showed a modestly increasing trend in site-wide abundance over time (D). Cerulean Warblers declined across the site over time (B), despite exhibiting an attraction to linear shale gas infrastructure. Linear trend lines (ordinary least squares re-gression) shown in blue.



Brown-headed Cowbirds were detected at only 3 points at the start of the study in 2008, but their abundance and spatial distribution increased across the site with the ex-pansion of shale gas (Figure 8). Low abundances of this species at the start of the study period led to reduced per-formance of interpolations, particularly in the early years of the study (Table 2). Nevertheless, interpolations for this species generally performed well, and indicate a clear pat-tern of attraction to areas disturbed by shale gas (Figure 8). This pattern is also supported by our distance-based

models, with greater abundances of Brown-headed Cowbirds near both well pads and linear infrastructure (Table 1), and overall increasing abundances across the site over time (Figure 5C).

DISCUSSION

More than half of the species we evaluated were sensi-tive to proximity to shale gas, and among forest interior birds we found more negative than positive effects. In

FIGURE 6. Annual interpolated distribution probability maps for Cerulean Warblers based on indicator kriging. GLMMs showed Cerulean Warblers had greater abundances in proximity to linear shale gas infrastructure (Table 1). In interpolated maps, darker red areas represent locations with greater probability of occurrence, and all shale gas is shown in black, for reference.



a previous study at this site, forest interior guild rich-ness decreased with increasing shale gas area within 100 m, and most forest interior focal species decreased in abundance with increasing shale gas area and forest edge density within 100 m (Farwell et al. 2016). Here we demonstrate that avoidance of shale gas exacerbated the negative effects of direct habitat loss on forest interior–dependent birds. Because avoidance extended well be-yond the area immediately disturbed by infrastructure (Figure 2A,2B), shale gas expansion dramatically reduced the amount of suitable interior forest habitat, altered

species distributions, and contributed to site-wide de-clines in forest interior birds.

Our finding that edge-avoiding, forest interior obli-gates decline in proximity to shale gas is consistent with other studies of energy development impacts on birds. In other parts of the central Appalachian region, forest in-terior birds decreased in abundance closer to unconven-tional shale gas well pads (Barton et al. 2016) and pipelines (Langlois 2017), and at sites containing conventional oil and gas wells (Thomas et al. 2014). In the boreal forests of western Canada, birds associated with older forests were

FIGURE 7. Annual interpolated distribution probability maps for Indigo Buntings based on indicator kriging. GLMMs showed Indigo Buntings occurred in greater abundances closer to shale gas linear infrastructure, in particular (Table 1). In interpolated maps, darker red areas represent locations with greater probability of occurrence, and all shale gas is shown in black, for reference.



less abundant near one or more types of conventional oil and gas infrastructure (Bayne et al. 2016). In particular, Ovenbirds showed negative responses to seismic lines in boreal forests (Bayne et al. 2005, Machtans 2006). In the western United States, sagebrush obligate songbirds de-clined with increasing densities of oil and gas well pads (Gilbert and Chalfoun 2011) and roads associated with natural gas development (Mutter et al. 2015). Similarly, in North American prairies, grassland obligate songbirds avoided unconventional shale oil well pads and roads

(Thompson et al. 2015), and declined with increasing nat-ural gas well density (Hamilton et al. 2011) and at sites con-taining oil infrastructure (Nenninger and Koper 2018).

We add to a growing body of research documenting negative effects of energy sector development on in-terior habitats and area-sensitive birds. However, it is important to note we observed considerable variation in species-specific responses, with some forest interior gap specialists (i.e. American Redstart, Cerulean Warbler, and Eastern Wood-Pewee) showing an attraction to shale

FIGURE 8. Annual interpolated distribution probability maps for Brown-headed Cowbirds based on indicator kriging. GLMMs showed Brown-headed Cowbirds occurred in greater abundances with increasing proximity to both shale gas well pads and shale gas linear infrastructure (Table 1). In interpolated maps, darker red areas represent locations with greater probability of occurrence, and all shale gas is shown in black, for reference.



gas. Forest interior gap-associated birds also increased in abundance with increasing forest edge density resulting from linear infrastructure at this study site (Farwell et al. 2016). In other eastern U.S. forests, Eastern Wood-Pewees were more abundant closer to shale gas well pads (Barton et al. 2016), and American Redstarts were more abun-dant within 50 m of shale gas pipelines than at reference points (Langlois 2017). Similarly, 5 species associated with older boreal forests were more abundant along pipelines or seismic lines than in undisturbed forests (Bayne et al. 2016). Relatively small forest openings created by oil and gas development may mimic natural canopy gaps and at-tract these species (Zmihorski 2012, Perkins and Wood 2014). However, the ecological effects of forest gaps cre-ated by anthropogenic vs. natural disturbances are likely to differ (Hobson et al. 1999, Imbeau et al. 1999, Vepakomma et al. 2018), depending on the extent to which anthropo-genic disturbance resembles the structure of surrounding habitat, and the size and severity of natural disturbances (Marzluff and Ewing 2001). Comparisons of avian response to anthropogenic vs. natural forest gaps have focused on timber harvests (Forsman et al. 2010). However, harvests begin to regenerate relatively quickly, and represent a form of human disturbance less dramatically different from sur-rounding forests than the hard edges, impervious surfaces, and functionally permanent disturbances typical of roads and energy development (Ortega and Capen 2002, Evans and Kiesecker 2014).

Our results also indicate that synanthropic species and some early successional species showed a spatial pattern of attraction to shale gas. However, we observed more synanthropic than early successional species with clear positive responses to well pads and linear infrastructure. Similarly, synanthropic birds increased in abundance with increasing shale gas area and forest edges at this site, whereas early successional birds only showed a positive re-sponse to forest edge density driven by expanding linear infrastructure (Farwell et al. 2016). Additional studies have documented a pattern of attraction among human-adapted species, and to a lesser extent among early successional species, to energy sector disturbances. In other central Appalachian forests, synanthropic species increased in

abundance and early successional species showed mixed responses near shale gas well pads (Barton et al. 2016) and pipelines (Langlois 2017), whereas species richness among both synanthropic and early successional guilds was higher at sites with conventional oil and gas wells (Thomas et al. 2014). These responses suggest areas disturbed by energy development represent habitat gains for human-adapted species and habitat generalists, while suitability of these areas for early successional species may depend on how closely development resembles natural disturbance.

Similar findings of “winners and losers” in human-modified landscapes (McKinney and Lockwood 1999) have been well documented for comparable forms of develop-ment, including roads (Fahrig and Rytwinski 2009, Benitez-Lopez et al. 2010), utility rights-of-way (Yahner et al. 2003, Richardson et al. 2017), urban and exurban encroachment (Suarez-Rubio et al. 2013, Rayner et al. 2015), mountaintop mining (Becker et al. 2015, Margenau et al. 2019), and re-newable energy development (Shafer and Buhl 2016, Smith and Dwyer 2016). Potential mechanisms driving these pat-terns of response include habitat loss (Villard et al. 1999, Mortelliti et al. 2010), decreased patch size (Boulinier et al. 2001, Davis 2004), increased edge creation (Banks-Leite et al. 2010), and greater rates of predation (Marzluff et al. 2004, Hethcoat and Chalfoun 2015), interspecific compe-tition (Nee and May 1992, Shochat et al. 2010), and nest parasitism (Gates and Evans 1998, Howell et al. 2007) in isolated patches and along edges. Brown-headed Cowbirds consistently occurred in greater numbers near energy de-velopment in studies we reviewed (Thomas et al. 2014, Thompson et al. 2015, Barton et al. 2016, Bayne et al. 2016, Farwell et al. 2016, Langlois 2017) and resulted in increased rates of nest parasitism near oil and gas development in the Northern Great Plains (Bernath-Plaisted et al. 2017) and at our study site in the central Appalachians (Frantz et al. 2018b).

Expanding infrastructure also increases human access and activity, including traffic (Ingelfinger and Anderson 2004), light (Kempenaers et al. 2010, de Jong et al. 2016), and noise pollution (Halfwerk et al. 2011, Ware et al. 2015). Human activities associated with oil and gas development in western Canada destroyed an estimated 10,200–41,150

TABLE 2. Root mean square of standardized error (RMSSE) values for maps interpolated with indicator kriging (Figures 4, 6, 7, 8). This measure of interpolation performance represents the relationship between measured and estimated variances, and should op-timally be 1 (Johnston et al. 2003). If the RMSSE value is >1, kriging variances are underestimated; if it is <1, they are overestimated. Interpolations with RMSSE values close to 1 (i.e. ±0.075; Polakowska et al. 2012) have valid prediction standard errors and can be con-sidered reliable (Johnston et al. 2003).

Species 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017

Ovenbird 1.000 0.981 0.982 1.001 0.960 0.986 0.987 1.007 0.996 1.018Cerulean Warbler 0.999 0.992 0.993 0.980 1.009 1.002 1.028 1.006 1.024 1.013Indigo Bunting 1.019 1.044 1.022 1.013 1.009 0.990 0.991 1.065 1.010 1.027Brown-headed Cowbird 1.201 1.152 1.023 1.098 1.003 1.061 1.011 1.024 1.068 1.078



avian nests annually (Van Wilgenburg et al. 2013), and Killdeer (Charadrius vociferous) nests found on gravel oil pads were more likely to be destroyed by machinery and human activity than nests in grazed pastures (Atuo et al. 2018). Noise impacts associated with oil and gas devel-opment may pose a particular problem for breeding birds that rely on acoustic communication for mate attraction, territory defense, and parent–offspring communication (Rabin et al. 2003). During active phases of construction and hydraulic fracturing, heavy machinery and high vol-umes of truck traffic create acute noise impacts lasting from months to years, depending on the number of wells (Brittingham et al. 2014). Pipeline compressor stations also operate constantly and are a source of long-term, chronic noise that can extend up to 700 m into surrounding habitat (Bayne et al. 2008, Francis et al. 2011). Although Bayne et al. (2008) found overall passerine density was lower near compressor stations, and Francis et al. (2009, 2011) found some species had lower occupancy and avoided nesting near compressor stations, these studies found highly vari-able species-specific responses. Some birds avoided noisy areas while others appeared to be more tolerant of noise; however, there is evidence that noise still increases physio-logical stress (Blickley et al. 2012) and reduces fitness (Kleist et al. 2018) in birds that remain in noisy areas.

Lastly, there is growing evidence that surface water contamination associated with hydraulic fracturing may negatively impact stream-dependent species. Streams impacted by shale gas development in the central Appalachian region had higher methyl mercury con-centrations, were more acidic, and had lower benthic macroinvertebrate diversity compared with unimpacted streams (Grant et al. 2016, Wood et al. 2016, Frantz et al. 2018a). The Louisiana Waterthrush, a forest song-bird that feeds primarily on benthic macroinvertebrates, bioaccumulated heavy metals associated with hydraulic fracturing in watersheds containing shale gas, but not in reference watersheds (Latta et al. 2015). Louisiana Waterthrushes also had lower rates of reproductive suc-cess in territories disturbed by shale gas than undisturbed territories at our long-term study site (Wood et al. 2016, Frantz et al. 2018b).

These findings underscore the importance of further research on avian demographic rates in areas affected by shale gas. Our analyses were limited to species abun-dance data, but species abundance is not always an ac-curate measure of habitat quality (Van Horne 1983). Forest interior gap specialists such as Cerulean Warblers occurred in higher numbers near shale gas, but still de-clined in overall abundance across our study area over time (Figure 5B; Farwell et al. 2016), which suggests shale gas infrastructure may be functioning as an ecological trap (Weldon and Haddad 2005, Boves et al. 2013b, Atuo et al. 2018). Cerulean Warblers frequently locate nest sites near

canopy gaps, which may increase their vulnerability to nest predation and parasitism near human-caused forest edges (Bakermans and Rodewald 2009, Boves et al. 2013a). This hypothesis is supported by increased abundances of Brown-headed Cowbirds near shale gas infrastructure and increased rates of nest parasitism of another forest interior species at this site (Frantz et al. 2018b). Finally, there is some evidence that bioaccumulation of toxins associated with hydraulic fracturing (Latta et al. 2015) and increased physiological stress from exposure to chronic noise (Kleist et al. 2018) and light pollution (Kempenaers et al. 2010, de Jong et al. 2016) may lower the fitness of birds that remain in areas impacted by shale gas infrastructure.

CONCLUSION

The unconventional shale gas industry represents a rela-tively recent form of energy development undergoing a rapid expansion, both regionally and globally (USEIA 2015, 2018). Here we add to a growing body of research indicating shale gas development is negatively affecting forest ecosystems and restructuring native biological communities. This is not a trivial issue, considering the Marcellus-Utica shale region holds one of the largest volumes of recoverable natural gas in North America and continues to drive U.S. shale gas production growth (USEIA 2018). A number of interstate pipelines are cur-rently under construction in the eastern United States (FERC 2018), many of which will bisect remaining areas of interior forest and spur additional development of new well pads, gathering lines, and access roads. Our results suggest efforts to avoid development of shale gas infra-structure in core forest areas of high conservation value will help mitigate negative impacts to native passerine communities in a region with high value for global bio-diversity (Gillen and Kiviat 2012, Kiviat 2013).

ACKNOWLEDGMENTS

West Virginia Division of Natural Resources granted access to the study area, Wheeling Jesuit University provided ac-cess to field housing, and West Virginia Cooperative Fish and Wildlife Research unit provided field vehicles, equipment, and logistic support. We thank the many graduate students and field surveyors who collected data over the 10-yr study. Special thanks to Michael Strager for input on study design and project feedback. James Anderson, Margaret Brittingham, Brenden McNeil, Cynthia Loftin, and 2 anonymous reviewers provided helpful comments on an earlier version of this manuscript. Any use of trade, firm, or product names is for descriptive purposes only and does not imply endorsement by the U.S. Government.Funding statement: This research was funded by the NETL Department of Energy, West Virginia Division of Natural Resources, and U.S. Fish and Wildlife Service.



Author contributions: L.S.F. and P.B.W. conceived the idea, design, and experiment (supervised research, formulated question or hypothesis). L.S.F., P.B.W., and D.J.B. developed or designed methods. L.S.F. and J.S. performed the experiments (collected data, conducted the research). L.S.F., D.J.B., and J.S. analyzed the data. L.S.F. and P.B.W. wrote the paper.

LITERATURE CITED

Abdi, A., and A. Nandipati (2009). Bird diversity modeling using geostatistics and GIS. In: 12th AGILE international conference on geographic information science, 2-5 June 2009. Hannover, Germany: Leibniz Universitat Hannover, 1–20.

Alldredge, M. W., T. R. Simons, and K. H. Pollock (2007). A field evalu-ation of distance measurement error in auditory avian point count surveys. Journal of Wildlife Management 71:2759–2766.

Atuo, F. A., P. Saud, C. Wyatt, B. Determan, J. A. Crose, and T. J. O’Connell (2018). Are oil and natural gas development sites ecological traps for nesting Killdeer? Wildlife Biology 1(476):1-8.

Bakermans, M. H., and A. D. Rodewald (2009). Think globally, manage locally: The importance of steady-state forest features for a declining songbird. Forest Ecology and Management 258:224–232.

Banks-Leite, C., R. M. Ewers, and J. P. Metzger (2010). Edge effects as the principal cause of area effects on birds in fragmented secondary forest. Oikos 119:918–926.

Barton, E. P., S. E. Pabian, and M. C. Brittingham (2016). Bird com-munity response to Marcellus shale gas development. Journal of Wildlife Management 80:1301–1313.

Bates, D., M. Maechler, B. Bolker, and S. Walker (2015). Fitting linear mixed-effects models using lme4. Journal of Statistical Software 67:1–48.

Bayne, E. M., and B. C. Dale (2011). Effects of energy develop-ment on songbirds. In Energy Development and Wildlife Conservation in Western North America (D. E. Naugle, Editor). Island Press, Washington, D.C., USA.

Bayne, E. M., L. Habib, and S. Boutin (2008). Impacts of chronic an-thropogenic noise from energy-sector development on abun-dance of songbirds in the boreal forest. Conservation Biology 22:1186–1193.

Bayne, E. M., L. Leston, C. L. Mahon, P. Solymos, C. Machtans, H. Lankau, J. R. Ball, S. L. Van Wilgenburg, S. G. Cumming, T. Fontaine, F. K. A. Schmiegelow, and S. J. Song (2016). Boreal bird abundance estimates within different energy sector disturbances vary with point count radius. The Condor: Ornithological Applications 118:376–390.

Bayne, E. M., S. L. Van Wilgenburg, S. Boutin, and K. A. Hobson (2005). Modeling and field-testing of Ovenbird (Seiurus aurocapillus) responses to boreal forest dissection by energy sector development at multiple spatial scales. Landscape Ecology 20:203–216.

Becker, D. A., P. B. Wood, M. P. Strager, and C. Mazzarella (2015). Impacts of mountaintop mining on terrestrial ecosystem in-tegrity: Identifying landscape thresholds for avian species in the central Appalachians, United States. Landscape Ecology 30:339–356.

Benítez-López, A., R. Alkemade, and P. A. Verweij (2010). The im-pacts of roads and other infrastructure on mammal and

bird populations: A meta-analysis. Biological Conservation 143:1307–1316.

Bernath-Plaisted, J., H. Nenninger, and N. Koper (2017). Conventional oil and natural gas infrastructure increases Brown-headed Cowbird (Molothrus ater) relative abundance and parasitism in mixed-grass prairie. Royal Society Open Science 4:170036.

Bibby, C. J., N. D. Burgess, and D. A. Hill (1992). Bird Census Techniques. Academic Press, London, UK.

Blickley, J. L., K. R. Word, A. H. Krakauer, J. L. Phillips, S. N. Sells, C. C. Taff, J. C. Wingfield, and G. L. Patricelli (2012). Experimental chronic noise is related to elevated fecal corticosteroid me-tabolites in lekking male Greater Sage-Grouse (Centrocercus urophasianus). PLOS One 7:e50462.

Boulinier, T., J. D. Nichols, J. E. Hines, J. R. Sauer, C. H. Flather, and K. H. Pollock (2001). Forest fragmentation and bird community dynamics: Inference at regional scales. Ecology 82:1159–1169.

Boves, T. J., D. A. Buehler, J. Sheehan, P. B. Wood, A. D. Rodewald, J. L. Larkin, P. D. Keyser, F. L. Newell, A. Evans, G. A. George, and T. B. Wigley (2013a). Spatial variation in breeding habitat selec-tion by Cerulean Warblers (Setophaga cerulea) throughout the Appalachian Mountains. The Auk 1:46–59.

Boves, T. J., D. A. Buehler, J. Sheehan, P. B. Wood, A. D. Rodewald, J. L. Larkin, P. D. Keyser, F. L. Newell, G. A. George, M. H. Bakermans, et al. (2013b). Emulating natural disturbances for declining late-successional species: A case study of the consequences for Cerulean Warblers (Setophaga cerulea). PLOS One 8:e52107.

Brittingham, M. C., K. O. Maloney, A. M. Farag, D. D. Harper, and Z. H. Bowen (2014). Ecological risks of shale oil and gas de-velopment to wildlife, aquatic resources and their habitats. Environmental Science & Technology 48:11034–11047.

Davis, S. K. (2004). Area sensitivity in grassland passerines: Effects of patch size, patch shape, and vegetation structure on bird abundance and occurrence in southern Saskatchewan. The Auk 121:1130–1145.

DeCapita, M. E. (2000). Brown-headed Cowbird control on Kirtland’s Warbler nesting areas in Michigan, 1972–1995. In Ecology and Management of Cowbirds and Their Hosts: Studies in the Conservation of North American Passerine Birds (J. N. M. Smith, T. L. Cook, S. I. Rothstein, S. K. Robinson, and S. G. Sealy, Editors). University of Texas Press, Austin, TX, USA. pp. 333–341.

de Jong, M., L. Jeninga, J. Q. Ouyang, K. van Oers, K. Spoelstra, and M. E. Visser (2016). Dose-dependent responses of avian daily rhythms to artificial light at night. Physiology & Behavior 155:172–179.

Drohan, P. J., M. Brittingham, J. Bishop, and K. Yoder (2012). Early trends in landcover change and forest fragmentation due to shale-gas development in Pennsylvania: A potential out-come for the northcentral Appalachians. Environmental Management 49:1061–1075.

Eckrich, G. H., J. A. Koloszar, and M. D. Goering (1999). Effective landscape management of Brown-headed Cowbirds at Fort Hood, Texas. In Research and Management of the Brown-headed Cowbird in Western Landscapes (M. L. Morrison, L. S. Hall, S. K. Robinson, S. I. Rothstein, D. Caldwell Hahn, and T. D. Rich, Editors). Studies in Avian Biology 18:267–274.

Ehrlich, P. R., D. S. Dobkin, and D. Wheye (1988). The Birder’s Handbook: A Field Guide to the Natural History of North American Birds. Simon & Schuster, New York, NY, USA.



[ESRI] Environmental Research Systems Institute (2016). ArcGIS Desktop: Release 10.4. Redlands, CA, USA.

Evans, J. S., and J. M. Kiesecker (2014). Shale gas, wind and water: Assessing the potential cumulative impacts of energy devel-opment on ecosystem services within the Marcellus play. PLOS One 9:e89210.

Fahrig, L., and T. Rytwinski (2009). Effects of roads on animal abun-dance: An empirical review and synthesis. Ecology and Society 14:art.21.

Farwell, L. S., P. B. Wood, J. Sheehan, and G. A. George (2016). Shale gas development effects on the songbird community in a cen-tral Appalachian forest. Biological Conservation 201:78–91.

[FERC] Federal Energy Regulatory Commission (2018). Approved major natural gas pipeline projects. https://www.ferc.gov/in-dustries/gas/indus-act/pipelines/approved-projects.asp

Forsman, J. T., P. Reunanen, J. Jokimäki, and M. Mönkkönen (2010). The effects of small-scale disturbance on forest birds: A meta-analysis. Canadian Journal of Forest Research 40: 1833–1842.

Francis, C. D., C. P. Ortega, and A. Cruz. 2009. Noise pollution changes avian communities and species interactions. Current Biology 19:1415–1419.

Francis, C. D., J. Paritsis, C. P. Ortega, and A. Cruz. 2011. Landscape patterns of avian habitat use and nest success are affected by chronic gaswell compressor noise. Landscape Ecology 26:1269–1280.

Frantz, M. W., P. B. Wood, and G. T. Merovich, Jr. (2018a). Demographic characteristics of an avian predator, Louisiana Waterthrush (Parkesia motacilla), in response to its aquatic prey in a central Appalachian USA watershed impacted by shale gas development. PLOS One 13:e0206077.

Frantz, M. W., P. B. Wood, J. Sheehan, and G. A. George (2018b). Demographic response of Louisiana Waterthrush, a stream ob-ligate songbird of conservation concern, to shale gas develop-ment. The Condor: Ornithological Applications 120:265–282.

Gaston, K. J., T. M. Blackburn, and K. K. Goldewijk (2003). Habitat conversion and global avian biodiversity loss. Proceedings of the Royal Society B 270:1293–1300.

Gates, J. E., and D. R. Evans (1998). Cowbirds breeding in the cen-tral Appalachians: Spatial and temporal patterns and habitat selection. Ecological Applications 8:27–40.

Gilbert, M. M., and A. D. Chalfoun (2011). Energy development af-fects populations of sagebrush songbirds in Wyoming. Journal of Wildlife Management 75:816–824.

Gillen, J. L., and E. Kiviat (2012). Hydraulic fracturing threats to spe-cies with restricted geographic ranges in the eastern United States. Environmental Practice 14:320–331.

Grant, C. J., A. K. Lutz, A. D. Kulig, and M. R. Stanton (2016). Fracked ecology: Response of aquatic trophic structure and mercury biomagnification dynamics in the Marcellus shale formation. Ecotoxicology 25:1739–1750.

Halfwerk, W., L. J. M. Holleman, C. M. Lessells, and H. Slabbekoom (2011). Negative impact of traffic noise on avian reproductive success. Journal of Applied Ecology 48:210–219.

Hamilton, L. E., B. C. Dale, and C. A. Paszkowski (2011). Effects of disturbance associated with natural gas extraction on the oc-currence of three grassland songbirds. Avian Conservation and Ecology 6(1):7.

Hansen, A. J., and J. J. Rotella (2000). Bird responses to forest fragmentation. In Fragmentation in the Southern Rocky Mountains (R. L. Knight, F. W. Smith, S. W. Buskirk, W. H. Romme,

and W. L. Baker, Editors). University Press of Colorado, Boulder, Colorado, USA. pp. 201–219.

Hethcoat, M. G., and A. D. Chalfoun (2015). Towards a mech-anistic understanding of human-induced rapid environ-mental change: A case study linking energy development, nest predation, and predators. Journal of Applied Ecology 52:1492–1499.

Hobson, K. A., and J. Schieck (1999). Changes in bird communities in boreal mixedwood forest: Harvest and wildfire effects over 30 years. Ecological Applications 9:849–863.

Howell, C. A., W. D. Dijak, and F. R. Thompson (2007). Landscape context and selection for forest edge by breeding Brown-headed Cowbirds. Landscape Ecology 22:273–284.

Imbeau, L., J. P. L. Savard, and R. Gagnon (1999). Comparing bird assemblages in successional black spruce stands originating from fire and logging. Canadian Journal of Zoology 77:1850–1860.

Ingelfinger, F., and S. Anderson (2004). Passerine response to roads associated with natural gas extraction in a sagebrush steppe habitat. Western North American Naturalist 64:385–395.

Jetz, W., D. S. Wilcove, and A. P. Dobson (2007). Projected impacts of climate and land-use change on the global biodiversity of birds. PLoS Biology 5:1211–1219.

Johnston, K., J. M. Ver Hoef, K. Krivoruchko, and N. Lucas (2003). Using ArcGIS Geostatistical Analyst. ESRI, Redlands, CA, USA.

Kempenaers, B., P. Borgström, P. Loës, E. Schlicht, and M. Valcu (2010). Artificial night lighting affects dawn song, extra-pair siring success, and lay date in songbirds. Current Biology 20:1735–1739.

Kiviat, E. (2013). Risks to biodiversity from hydraulic fracturing for natural gas in the Marcellus and Utica shales. Annals of the New York Academy of Sciences 1286:1–14.

Kleist, N. J., R. P. Guralnick, A. Cruz, C. A. Lowry, and C. D. Francis (2018). Chronic anthropogenic noise disrupts glucocorticoid signaling and has multiple effects on fitness in an avian com-munity. Proceedings of the National Academy of Sciences USA 115:648–657.

Kus, B. E. (1999). Impacts of Brown-headed Cowbird para-sitism on productivity of the endangered Least Bell’s Vireo. In Research and Management of the Brown-headed Cowbird in Western Landscapes (M. L. Morrison, L. S. Hall, S. K. Robinson, S. I. Rothstein, D. Caldwell Hahn, and T. D. Rich, Editors). Studies in Avian Biology 18:160–166.

Langlois, L. A. (2017). Effects of Marcellus shale gas infrastructure on forest fragmentation and bird communities in northcentral Pennsylvania. Ph.D. dissertation, Pennsylvania State University, State College, PA, USA.

Langlois, L. A., P. J. Drohan, and M. C. Brittingham (2017). Linear infrastructure drives habitat conversion and forest fragmen-tation associated with Marcellus shale gas development in a forested landscape. Journal of Environmental Management 197:167–176.

Latta, S. C., L. C. Marshall, M. W. Frantz, and J. D. Toms (2015). Evidence from two shale regions that a riparian songbird accumulates metals associated with hydraulic fracturing. Ecosphere 6:art.144.

Li, J., and A. D. Heap (2008). A Review of Spatial Interpolation Methods for Environmental Scientists. Geoscience Australia, Record 2008/23.

Lichstein, J. W., T. R. Simons, and K. E. Franzreb (2002). Landscape effects on breeding songbird abundance in managed forests. Ecological Applications 12:836–857.

https://www.ferc.gov/industries/gas/indus-act/pipelines/approved-projects.asp

https://www.ferc.gov/industries/gas/indus-act/pipelines/approved-projects.asp



Lüdecke, D. (2019). sjstats: Statistical functions for regression models. R package version 0.17.3.

Machtans, C. S. (2006). Songbird response to seismic lines in the western boreal forest: A manipulative experiment. Canadian Journal of Zoology 84:1421–1430.

Margenau, E. L., P. B. Wood, C. A. Weakland, and D. J. Brown (2019). Trade-offs relating to grassland and forest mine reclamation approaches in the central Appalachian region and its impli-cations for the songbird community. Avian Conservation and Ecology 14:art.2.

Marzluff, J. M., and K. Ewing (2001). Restoration of fragmented landscapes for the conservation of birds: A general framework and specific recommendations for urbanizing landscapes. Restoration Ecology 9:280–292.

Marzluff, J. M., J. J. Millspaugh, P. Hurvitz, and M. S. Handcock (2004). Relating resources to a probabilistic measure of space use: Forest fragments and Steller’s Jays. Ecology 85:1411–1427.

McGarigal, K., S. A. Cushman, and E. Ene (2012). FRAGSTATS v4: Spatial Pattern Analysis Program for Categorical and Continuous Maps. University of Massachusetts, Amherst, MA, USA. http://www.umass.edu/landeco/research/fragstats/fragstats.html

McKinney, M. L., and J. L. Lockwood (1999). Biotic homogeniza-tion: A few winners replacing many losers in the next mass ex-tinction. Trends in Ecology & Evolution 14:450–453.

Melles, S. J., M-J. Fortin, K. Lindsay, and D. Badzinski (2011). Expanding northward: influence of climate change, forest connectivity, and population processes on a threat-ened species’ range shift. Global Change Biology 17: 17–31.

Mortelliti, A., S. Fagiana, and C. Battisti (2010). Independent effects of habitat loss, habitat fragmentation and structural connect-ivity on forest-dependent birds. Diversity and Distributions. 16:941–951.

Mutter, M., D. C. Pavlacky, Jr., N. J. Van Lanen, and R. Grenyer (2015). Evaluating the impact of gas extraction infrastructure on the occupancy of sagebrush-obligate songbirds. Ecological Applications 25:1175–1186.

Nakagawa, S., P. C. D. Johnson, and H. Schielzeth (2017). The coefficient of determination R2 and intra-class correlation coefficient from generalized linear mixed effects models re-visited and expanded. Journal of the Royal Society Interface 14:art.20170213.

Nee, S., and R. M. May (1992). Dynamics of metapopulations: Habitat destruction and competitive coexistence. Journal of Animal Ecology 61:37–40.

Nenninger, H. R., and N. Koper (2018). Effects of conventional oil wells on grassland songbird abundance are caused by pres-ence of infrastructure, not noise. Biological Conservation 218:124–133.

Ortega, Y. K., and D. E. Capen (2002). Roads as edges: Effects on birds in forested landscapes. Forest Science 48:381–390.

Perkins, K. A., and P. B. Wood (2014). Selection of forest canopy gaps by male Cerulean Warblers in West Virginia. Wilson Journal of Ornithology 126:288–297.

Polakowska, A. E., M-J. Fortin, and A. Couturier (2012). Quantifying the spatial relationship between bird species’ distributions and landscape feature boundaries in southern Ontario, Canada. Landscape Ecology 27:1481–1493.

R Core Team (2017). R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing, Vienna, Austria. http://www/Rproject.org/

Rabin, L. A., B. McCowan, S. L. Hooper, and D. H. Owings (2003). Anthropogenic noise and its effect on animal communication: An interface between comparative psychology and conserva-tion biology. International Journal of Comparative Psychology 16:172–192.

Ralph, C. J., S. Droege, and J. R. Sauer (1995). Managing and moni-toring birds using point counts: Standards and applications. In: Monitoring Bird Populations by Point Counts (C. J. Ralph, J. R. Sauer, and S. Droege, Editors). U.S. Department of Agriculture, Forest Service, Pacific Southwest Research Station, General Technical Report PSW-GTR-149, Albany, CA, USA. pp. 161–168.

Rayner, L., K. Ikin, M. J. Evans, P. Gibbons, D. B. Lindenmayer, and A. D. Manning (2015). Avifauna and urban encroachment in time and space. Diversity and Distributions 21:428–440.

Richardson, M. L., B. A. Wilson, D. A. S. Aiuto, J. E. Crosby, A. Alonso, F. Dallmeier, and G. K. Golinski (2017). A review of the impact of pipelines and power lines on biodiversity and strategies for mitigation. Biodiversity and Conservation 26: 1801–1815.

Rodewald, P. G., Editor (2015). The Birds of North America. Cornell Lab of Ornithology, Ithaca, NY, USA.

Schmiegelow, F. K. A., C. S. Machtans, and S. J. Hannon (1997). Are boreal birds resilient to forest fragmentation? An experi-mental study of short-term community responses. Ecology 78:1914–1932.

Shafer, J. A., and D. A. Buhl (2016). Effects of wind-energy facil-ities on breeding grassland bird distributions. Conservation Biology 30:59–71.

Sheehan, J., P. B. Wood, D. A. Buehler, P. D. Keyser, and J. L. Larkin (2013). Avian response to timber harvesting applied experi-mentally to manage Cerulean Warbler breeding populations. Forest Ecology and Management 321:5–18.

Shochat, E., S. B. Lerman, J. M. Anderies, P. S. Warren, S. H. Faeth, and C. H. Nilon (2010). Invasion, competition, and biodiversity loss in urban ecosystems. Bioscience 60:199–208.

Smith, J. A., and J. F. Dwyer (2016). Avian interactions with re-newable energy infrastructure: An update. The Condor: Ornithological Applications 118:411–423.

Smith, E. P., D. R. Orvos, and J. Cairns. 1993. Impact assessment using the before-after-control-impact (BACI) model: Concerns and comments. Canadian Journal of Fisheries and Aquatic Sciences 50:627–637.

Sólymos, P., S. M. Matsuoka, E. M. Bayne, S. R. Lele, P. Fontaine, S. G. Cummings, D. Stralberg, F. K. A. Schmiegelow, and S. J. Song (2013). Calibrating indices of avian density from non-standardized survey data: Making the most of a messy situ-ation. Methods in Ecology and Evolution 4:1047–1058.

Solymos, P., M. Moreno, and S. R. Lele (2018). Analyzing wildlife data with detection error. R package version 0.4–2.

Suarez-Rubio, M., S. Wilson, P. Leimgruber, and T. Lookingbill (2013). Threshold responses of forest birds to land-scape changes around exurban development. PLOS One 8:e67593.

Thomas, E. H., M. C. Brittingham, and S. H. Stoleson (2014). Conventional oil and gas development alters forest songbird communities. Journal of Wildlife Management 78:293–306.

Thompson, S. J., D. H. Johnson, N. D. Niemut, and C. A. Ribic (2015). Avoidance of unconventional oil wells and roads exacerbates habitat loss for grassland birds in the North American great plains. Biological Conservation 192:82–90.

http://www.umass.edu/landeco/research/fragstats/fragstats.html

http://www.umass.edu/landeco/research/fragstats/fragstats.html

http://www/Rproject.org/



[USEIA] U.S. Energy Information Administration (2015). World shale resource assessments. https://www.eia.gov/analysis/studies/worldshalegas/

[USEIA] U.S. Energy Information Administration (2018). Annual energy outlook. https://www.eia.gov/outlooks/aeo/

Van Horne, B. (1983). Density as a misleading indicator of habitat quality. Journal of Wildlife Management 47:893–901.

Van Wilgenburg, S. L., K. A. Hobson, E. M. Bayne, and N. Koper (2013). Estimated avian nest loss associated with oil and gas exploration and extraction in the Western Canadian Sedimentary Basin. Avian Conservation and Ecology 8:art.9.

Vepakomma, U., D. D. Kneeshaw, and L. De Grandpré (2018). Influence of natural and anthropogenic linear canopy open-ings on forest structural patterns investigated using lidar. Forests 9:art.540.

Villard, M., M. K. Trzcinski, and G. Merriam (1999). Fragmentation effects on forest birds: Relative influence of woodland cover and configuration on landscape occupancy. Conservation Biology 13:774–783.

Walker, J. S., R. C. Balling, J. M. Briggs, M. Katti, P. S. Warren, and E. A. Wentz (2008). Birds of a feather: Interpolating distribution patterns of urban birds. Computers, Environment and Urban Systems 32:19–28.

Ware, H. E., C. J. W. McClure, J. D. Carlisle, and J. R. Barber (2015). A phantom road experiment reveals traffic noise is an invisible

source of habitat degradation. Proceedings of the National Academy of Sciences USA 112:12105–12109.

Weldon, A. J., and N. M. Haddad (2005). The effects of patch shape on Indigo Buntings: Evidence for an ecological trap. Ecology 86:1422–1431.

Wood, P. B., S. Bosworth, and R. Dettmers (2006). Cerulean Warbler abundance and occurrence relative to large-scale edge and habitat characteristics. The Condor 108:154–165.

Wood, P. B., M. W. Frantz, and D. A. Becker (2016). Louisiana Waterthrush and benthic macroinvertebrate response to shale gas development. Journal of Fish and Wildlife Management 7:423–433.

Yahner, R. H., R. J. Hutnik, and S. A. Liscinsky (2003). Long-term trends in bird populations on an electric transmission right-of-way. Journal of Arboriculture 29:156–161.

Zmihorski, M. (2012). The effects of anthropogenic and natural disturbances on breeding birds of managed Scots pine forests in northern Poland. Ornis Fennica 89: 63–73.

Zuur, A. F., E. N. Ieno, and C. S. Elphick (2010). A protocol for data exploration to avoid common statistical problems. Methods in Ecology and Evolution 1:3–14.

Zuur, A. F., E. N. Ieno, N. J. Walker, A. A. Saveliev, and G. M. Smith (2009). Mixed Effects Models and Extensions in Ecology with R. Springer-Verlag, New York, NY, USA.

APPENDIX TABLE 3. Sources of imagery for manual digitization of forest and non-forest cover.

Year Platform Season Resolution Source

2003 Aerial Leaf-off 0.61 m West Virginia Statewide Addressing and Mapping Board project2007 Aerial Leaf-on 1 m USDA National Agricultural Imagery Project2008 Satellite Leaf-on 30 m Landsat 72009 Satellite Leaf-on 0.65 m Quickbird2010 Satellite Leaf-on 30 m Landsat 72011 Aerial Leaf-on 1 m USDA National Agricultural Imagery Project2014 Aerial Leaf-on 1 m USDA National Agricultural Imagery Project2016 Aerial Leaf-on 1 m USDA National Agricultural Imagery Project

APPENDIX TABLE 4. Model sets included in AIC analysis of factors influencing detection probabilities using (A) time removal and (B) distance sampling. Models include intercept-only models as well as models incorporating survey-specific covariates. Observer experi-ence level (ranking of 1–5) was based on relative number of years of experience conducting avian point count surveys. Tree cover is percent forest cover within a 100-m radius of each survey point.

A. Time-removal models

Intercept-onlyTime since sunriseTime since sunrise + (Time since sunrise)2

Ordinal dateOrdinal date + (Ordinal date)2

Ordinal date + Time since sunriseOrdinal date + (Ordinal date)2 + Time since sunriseOrdinal date + Time since sunrise + (Time since sunrise)2

Ordinal date + (Ordinal date)2 + Time since sunrise + (Time since sunrise)2

B. Distance sampling models

Intercept-onlyTree cover (within 100 m)Observer experience levelObserver experience level + Tree cover (within 100 m)

https://www.eia.gov/analysis/studies/worldshalegas/

https://www.eia.gov/analysis/studies/worldshalegas/

https://www.eia.gov/outlooks/aeo/



APPENDIX TABLE 5. Cross validation of interpolated models (2008–2017) comparing performance of 3 interpolation methods: in-verse distance weighting (IDW), ordinary kriging, and indicator kriging (Johnston et al. 2003). Comparison is based on root mean square errors, with error values closer to zero indicating better fit between estimated and measured values.

Year IDW Ordinary Kriging Indicator Kriging

Ovenbird 2008 0.650 0.633 0.475 2009 0.622 0.609 0.468 2010 0.635 0.609 0.474 2011 0.564 0.558 0.449 2012 0.506 0.511 0.454 2013 0.586 0.575 0.456 2014 0.432 0.423 0.390 2015 0.540 0.551 0.423 2016 0.443 0.433 0.411 2017 0.509 0.498 0.402Cerulean Warbler 2008 0.992 1.002 0.483 2009 1.063 1.054 0.479 2010 0.906 0.859 0.481 2011 0.785 0.753 0.447 2012 0.767 0.699 0.482 2013 0.857 0.745 0.494 2014 0.725 0.519 0.426 2015 0.979 0.689 0.470 2016 0.500 0.498 0.408 2017 0.862 0.825 0.432Indigo Bunting 2008 0.979 0.986 0.438 2009 1.110 1.102 0.471 2010 0.982 0.996 0.469 2011 0.780 0.803 0.454 2012 0.590 0.774 0.476 2013 0.587 0.867 0.451 2014 0.511 0.753 0.417 2015 0.643 0.982 0.464 2016 0.882 0.841 0.427 2017 1.234 1.282 0.472Brown-headed Cowbird 2008 0.263 0.254 0.151 2009 0.384 0.394 0.210 2010 0.468 0.446 0.231 2011 0.834 0.878 0.296 2012 0.845 0.880 0.287 2013 2.796 2.965 0.437 2014 1.051 1.054 0.365 2015 2.772 2.871 0.415 2016 1.530 1.581 0.368 2017 1.234 1.275 0.341

proximity to unconventional shale gas infrastructure

Documents