Effects of Rotational Grazing on Grassland Songbirds on U.S. Dairy Farms

A THESIS SUBMITTED TO THE FACULTY OF THE GRADUATE SCHOOL

OF THE UNIVERSITY OF MINNESOTA BY

Kathryn Marie Clower

IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE

Nicholas P. Jordan and Todd W. Arnold

April 2011

© Kathryn Marie Clower 2011

i

Acknowledgements Credit is first due to the farmers who took an interest in this project and generously gave

us their time and access to their land. Enormous thanks goes to my advisors, Todd

Arnold and Nick Jordan, who offered endless amounts of time, guidance and

encouragement. Thanks to Laura Musacchio, who served on my committee and provided

essential input and advice. I would also like to thank the research team for providing me

with this opportunity and mentoring me through the research process: Kristen Nelson,

Steve Manson, and Bruce Vondrocek, with major assistance from Adam Berland, Rachel

Brummel, Dudley Bonsal, Genevieve Brand and Jenny Immich. Thanks to our amazing

field crew, Sarah Graves, Andrew Nessel and Sondra Campbell, and to Sheri Huerd for

logistical assistance. I would like to thank Mark Holland and the University of

Minnesota Statistical Consulting Clinic for a great deal of statistical assistance. My lab

mates, Sarah Thompson, Courtney Amundson and Brandt Meixell, provided feedback

and support in many areas of this project. Editing help came from Alexandra Swanson,

Seth Stapleton, Jennifer Walton and other dear friends. Last but not least, I appreciate the

wonderful support provided by friends and family, especially Georgene Clower, April

Kaisen, and Gretchen Strobel.

ii

Dedication

This thesis is dedicated to the many wonderful teachers I’ve had,

both in and out of school. Most especially…

Ted Wieseman, for introducing me to natural history.

Charity Maybury, for the inspiration to go back to school.

iii

Abstract

Dairy pastures and hayfields provide important habitat for several grassland bird species

of conservation concern, and farm management can have a substantial impact on

populations of these species. Understanding the value of rotational grazing to grassland

birds may reveal new opportunities for bird conservation, and may clarify the degree to

which this management practice can enhance agricultural sustainability. We evaluated

management practices (e.g. grazing intensity) and habitat variables (e.g. land cover) at

100m and 1200m radii on 53 dairy farms in Wisconsin, Pennsylvania and New York, in

order to assess whether rotational grazing farms supported substantially higher

abundances of three grassland bird species – Savannah Sparrow, Bobolink, and

Grasshopper Sparrow -- than other dairy farms. We conducted 2-way ANOVAs in SAS

(9.2) to examine differences in relative abundance among states and farm types using

PROC GENMOD. Additionally, we modeled relative abundance as a response to Julian

date and land cover at 100m and 1200m radii using PROC GLIMMIX. Abundances of

each grassland bird species did not differ significantly among farm types. Responses to

land-cover variables differed among species but demonstrated that relative abundance

was positively associated with the proportion of hayfields and pastures in the surrounding

landscape, and negatively associated with woody cover. Savannah Sparrows and

Grasshopper Sparrows each showed a significant response to at least one habitat variable

at the 1200m scale. Our results suggest that differences in abundance of grassland birds

among farm types are modest at best, and that the impact of management practices on

these species must be understood in the context of the surrounding landscape. Our data

iv

show that grassland birds can benefit from increases in the total area of pasture and

hayfields on the landscape, which could be provided through broad adoption of rotational

grazing. However, the benefits provided by increasing grassland habitats on a small

number of farms are unlikely to have a substantial impact on bird populations,

particularly in the context of a highly fragmented agricultural landscape. Future research

for conservation planning and policy development should focus on the landscape scale

(i.e. ! 500 ha) to ensure that conservation actions are most effective.

v

Table of Contents

1. List of Tables ................................................................................................................. vi

2. List of Figures ............................................................................................................... vii

3. Chapter One. Multifunctionality: For the birds?..............................................................1

4. Chapter Two. The effects of rotational grazing on relative abundance of grassland-

obligate songbirds on U.S. dairy farms. ..........................................................................8

a. Introduction .................................................................................................................8

b. Methods .....................................................................................................................10

c. Results .......................................................................................................................15

d. Discussion .................................................................................................................17

e. Management Implications .........................................................................................20

5. Literature Cited ..............................................................................................................22

6. Tables.............................................................................................................................28

7. Figures............................................................................................................................34

vi

List of Tables

Table 1: Description, mean and range of values for explanatory variables used to model

relative abundance of grassland birds ...........................................................................28

Table 2: Acreage and land cover distribution on study farms, by farm type and by state.29

Table 3: Bird species observed on at least 14 of the 53 study farms .................................30

Table 4: Mean observed abundance per point-count circle (3.14 ha) of grassland bird

species most commonly observed on study farms. .......................................................31

Table 5: Least-squares means of relative abundance.........................................................32

Table 6: Best supported models of relative abundance of grassland birds in relation to

land-cover variables at 100m and 1200m radii .............................................................33

vii

List of Figures

Figure 1: Locations of counties (shaded) containing study sites in Wisconsin,

Pennsylvania and New York. ........................................................................................34

Figure 2. Landscape views of Clark County, WI (A), Lancaster County, PA (B) and

Madison County, NY (C). ............................................................................................35

Figure 3: Percent distribution of farm types surveyed in Wisconsin, Pennsylvania and

New York. .....................................................................................................................36

Figure 4: Aerial photo of a study farm, showing bird survey points (yellow) with 100m

(red) and 1200m (blue) radii. Farm fields are outlined in orange. ...............................37

Figure 5: Least-squares means and standard errors of bird abundance in response to

significant habitat variables, for Savannah Sparrow (A), Bobolink (B) and

Grasshopper Sparrow (C). .............................................................................................38

1

CHAPTER ONE

Mulfunctionality: For the birds?

Grassland bird populations across the United States have been declining since

surveys began in the 1960’s (Herkert 1995; Peterjohn and Sauer 1999; Vickery and

Herkert 2001). Although the factors responsible for these declines are varied and

complex, most scholars agree that a major factor is the loss and degradation of grassland

habitat, particularly due to agricultural practices (Bollinger et al. 1990; Murphy 2003;

Brennan and Kuvlesky 2005). Interestingly, the total extent of farmland has decreased

over the last 50 years (USDA Census of Agriculture). What has increased is the intensity

of farming practices on that land, including heavier application of synthetic fertilizers and

pesticides, earlier and more frequent mowing of hay, and homogenization of the

landscape through crop specialization (Askins et al. 2007). Grassland birds are known to

breed in surrogate grassland habitats, including farm pastures and hayfields (Paine et al.

1996; Askins et al. 2007), and are thus directly impacted by farming practices on these

landscapes. Any efforts to conserve these declining bird populations must therefore

address the management of working landscapes.

Working landscapes are defined as areas used primarily for agriculture, forestry or

fisheries production, and which may also support related industries such as tourism,

recreation, mining, energy production and other natural resource-based industries (Morse

2010). While a strong argument can be made that all landscapes (agricultural, “natural”,

urban or otherwise) are shaped by feedback between human and natural systems, working

landscapes provide a particularly clear example of this eco-social feedback, and have

historically been marginalized in conservation efforts. Liu et al. (2007) define such

landscapes as “coupled human and natural systems”, wherein the cultural and bio-

physical landscapes are constantly co-creating each other (Naveh 1994; Bignal and

McCracken 1996; Liu et al. 2007). Successful conservation, of grassland birds as well as

other natural resources, will require site-specific knowledge of eco-social feedback loops

and an understanding of how to influence these loops to enhance sustainability.

2

My thesis addresses the issue of sustainability on working lands at multiple levels.

The primary focus is an assessment of the impact of a particular pasture management

strategy, rotational grazing, on the presence and abundance of grassland passerines on

small U.S. dairy farms. However, this assessment comprises one portion of a much

broader interdisciplinary research project that used an evaluation of grazing practices on

dairy farms to understand eco-social feedback in multifunctional agriculture (MFA), with

related implications for agricultural sustainability at the landscape level. In this chapter, I

provide an overview of MFA and a description of the research project as a whole before

addressing the relationship of MFA to rotational grazing and grassland bird conservation.

Chapter 2 describes my research on grassland bird abundance in relation to dairy farm

management and land cover variables on the study farms.

Multifunctional Agriculture

The term “multifunctional agriculture” (MFA) was introduced at the Rio Earth

Summit in 1992 (Renting et al. 2009) and refers to the capacity for agricultural systems to

serve multiple functions beyond the production of commodities such as food and fiber

(Bills and Gross 2005; Vejre et al. 2007). Specifically, agricultural activities are directly

linked to both ecosystem functions and socio-economic functions. The former includes

environmental services such as biodiversity conservation, pest control and natural

resource management, while the latter involves agriculture and trade policy, rural

economic viability and preservation of cultural knowledge. MFA is therefore

interdisciplinary, highlighting the connections between agronomy, economics, ecology,

politics, and numerous other scientific and social fields. It is also multi-scalar, as

biophysical and social processes can be viewed from a local scale (e.g. individual fields

or farms), to a regional or even national scale (e.g. regional landscapes and climate, or

national farm policies). Note that the term “landscape” is used here to denote both the

biophysical (e.g. topography, biota, climate) and human (e.g. institutional, cultural, socio-

economic) realms, as well as the interactions between them (Naveh 1994).

Multifunctionality can be viewed as a measurable quality of a landscape. Some

agricultural systems may be more or less multifunctional than others, depending on the

3

variety of functions and benefits they provide. The more multifunctional a farm is, the

more resilient it will be – both to natural disturbances such as pests or weather events, as

well as to changes in economic markets and agricultural policies. In contrast to

industrialized agriculture that focuses exclusively on low-cost commodity production,

MFA has potential to provide increased non-market environmental and socio-economic

benefits (Boody et al. 2005; Jordan and Warner 2010). One way that this occurs is

through the use of management practices that employ eco-social feedback to enhance

ecological and socio-economic sustainability. In other words, in MFA, farming practices

and the human institutions that support them are intentionally brought into alignment

with ecosystem functions such that each supports the other.

Selman and Knight (2006) describe how such eco-social feedback can create a

“virtuous circle” that ensures sustainability, writing that “… in sustainable cultural

landscapes, people and place interact in mutually reinforcing ways, so that human agency

nurtures the resource base, which in turn rewards this stewardship by providing life-

support functions and economic opportunity.” However, human activities in the

industrialized world (including intensive agricultural practices) are increasingly

misaligned with natural systems, leading to widespread cultural and environmental

degradation (Naveh 1994; Selman and Knight 2006). “Landscape distinctiveness and

functionality are lost when this link breaks down,” write Selman and Knight, “so that

localities and regions become homogenized by ubiquitous trends, and scant social capital

is invested in cherishing and stewarding fragile places. In other words, the formerly

virtuous circle becomes a vicious circle… and there appears to be no spontaneous

mechanism whereby this process can be reversed: public intervention of some kind

appears necessary to initiate a process of re-embedding virtuosity” (Selman and Knight

2006: 298).

The interdisciplinary research project from which this thesis developed examined

whether one particular farming practice, rotational grazing, could be used to re-embed

virtuosity in landscapes that support the dairy industry. Our primary research question

was whether the use of rotational grazing (RG) led to measurable increases in

4

multifunctionality (and therefore sustainability) within a farm or region – and if so, how

did RG increase multifunctionality, and how much?

Rotational Grazing

Rotational grazing is a method of pasture management in which a grazing herd is

periodically rotated through small sections of pasture, called paddocks. This practice

prevents any single section from being overgrazed, and the alternating periods of

herbivory stimulate the grass to continue growing throughout the grazing season so that

the herd receives a constant source of nutritious feedstock. At the same time, the

presence of perennial cover on the landscape can improve soil structure and water

quality, while minimizing inputs of labor and other resources that would be required to

maintain annual crops (Undersander et al. 2002). RG is not a new technique, but has

gained increased attention since the 1990s as pressures in the dairy industry have led

many farmers to either expand and intensify their operations or to stay small and

specialize (Nott 2003; PATS 2007; Winsten et al. 2010). Estimates of current usage of

RG range from approximately 11-22% of Northeastern and 20-26% of Midwestern dairy

farms (Nott 2003; PATS 2007; Winsten et al. 2010). Less intensive forms of grazing are

used on an additional 20-50% of dairy farms (Agricultural Technology and Family Farm

Institute 1996; Gillespie et al. 2009); however, on these “continuous” or “low-intensity”

grazing farms, pasture is used primarily for exercise rather than nutrition, and the herd is

typically rotated infrequently if at all (Gillespie et al. 2009).

Previous studies suggest that rotational grazing is a form of MFA because it has

the potential to create multiple benefits beyond milk production, including socio-

economic benefits such as profitability and job satisfaction (Taylor and Foltz 2006) and

environmental benefits such as reduced erosion, increased wildlife habitat, animal

welfare, and improved animal and human health (Undersander et al. 2002). For example,

farms using RG can be at least as profitable as those using confinement methods (Kriegl

and McNair 2005; Gillespie et al. 2009) and many farmers who switch from

confinement-based dairy to a pasture-based system report that they have increased leisure

time and higher quality of life (Aschmann and Cropper 2007). Additionally, the use of

5

RG in riparian areas may be an effective method of rehabilitating streams (Lyons et al.

2000) and improving water quality (Sovell et al. 2000).

The MFA Research Project

In order to assess the extent to which the use of RG increases multifunctionality

on a landscape, our research team conducted social and biophysical surveys of a broad

sample of dairy farms across three states: Wisconsin, Pennsylvania and New York

(Figure 1). Our study farms used a variety of management strategies, ranging from full

confinement operations to intensively managed rotational grazing systems. The specific

methods we used to assess various biophysical and social components of our study farms

will be discussed in future publications, and most are not described here in detail. In

brief, the methods used in this project included qualitative interviews with farmers and

other key informants, a quantitative survey of farm and farmer demographics, biophysical

assessment of terrestrial and aquatic systems at multiple scales (including surveys of

streams, pasture plants, erosion hotspots, semi-natural areas, and farmland birds), digital

mapping and remote sensing of social and ecological phenomena, and integrated

modeling. We compared our findings across farms to determine whether use of RG was

associated with increased environmental and/or socio-economic benefits on specific

farms and within the local farming community.

Grassland Bird Abundance

Within this larger study of multifunctionality on dairy farms, my thesis work

focused on relative abundance of grassland birds as an indicator of the potential

environmental benefits generated by rotational grazing. Dairy pastures and hayfields

provide important habitat for several grassland bird species of conservation concern

(Paine et al. 1996; Askins et al. 2007), and pasture management may have a substantial

impact on populations of these species. My research questions were: Do grassland birds

occur in substantially greater abundance on dairy farms utilizing rotational grazing than

on other dairy farms? What mechanisms are responsible for observed differences in bird

abundance, and how are these mechanisms affected by management practices? Finally,

6

can management practices on a minority of farms affect grassland bird populations on a

regional scale? Understanding the value of RG to grassland birds may reveal new

opportunities for bird conservation, as well as clarifying the degree to which RG can

create virtuous circles and thereby enhance sustainability in agricultural landscapes.

Because of the complex relationship between management practices, habitat

features and bird abundance, the specific effects on grassland birds of rotational grazing

on dairy farms are not yet well understood. Although a number of studies have addressed

the impact of grazing on grassland birds, the majority of these have focused on managed

prairies or rangeland (typically for beef cattle) rather than the dairy landscape. Taken

collectively, these studies indicate that grazing can have both beneficial and detrimental

effects on birds. Paine et al. (1996) demonstrated that grazing intensity matters; nest

trampling on study pastures increased with length of the grazing rotation, with minimal

trampling observed on dairy pastures where the grazing herd was rotated daily. Walk and

Warner (2000) assert that some grazing systems can create the type of heterogeneous

landscape that is attractive to grassland birds, but that individual bird species show

different responses to various levels of grazing intensity. Jobin et al. (1998) suggest that

dairy farms, regardless of grazing intensity, may support higher abundances of grassland

birds than cash-crop farms because of the increased vegetative cover and decreased

intensity of management on dairy farms. These studies provide insight into the effects of

grazing on grassland birds, both its direct impacts (e.g. nest trampling by grazing cows)

and its indirect impacts (e.g. changes to vegetation structure that may affect territory

selection, vulnerability to nest predation, etc.). However, none of the studies clearly

confirm or deny that adoption of rotational grazing on dairy farms is likely to have

meaningful benefits to grassland bird populations. Furthermore, few address the

combined effects of grazing intensity and habitat factors at more than one spatial scale.

Our study examined the effects of both management (i.e. grazing intensity) and

habitat factors (i.e. land cover) impacting grassland bird densities on dairy farms and

assessed the relative value of RG in the context of habitat variables at two scales. Our

methods allowed us to compare bird abundance among a range of farm types while

accounting for the land cover context of both the immediate breeding territory (100m

7

radius from survey points) and the surrounding landscape (1200m radius). By shedding

light on the potential of RG to support grassland bird populations, this combined look at

management practices, landscape factors, and bird abundance has value to

conservationists as well as proponents of MFA.

The remainder of my thesis is formatted for The Wilson Journal of Ornithology.

Because this chapter will be jointly submitted with my advisors as coauthors, I have used

plural pronouns throughout the remainder of Chapter 2. However, this thesis represents

my own work and I take full responsibility for any errors or omissions.

8

CHAPTER TWO

The impact of rotational grazing on relative abundance of grassland-obligate songbirds

on U.S. dairy farms

INTRODUCTION

Declines in populations of North American grassland birds are well documented

(Herkert 1995; Peterjohn and Sauer 1999; Murphy 2003; Brennan and Kuvleski 2005;

Askins et al. 2007) and a growing body of literature explores the various causes and

potential management responses to these declines (Bollinger et al. 1990; Herkert 1997;

Helzer and Jelinski 1999; Vickery and Herkert 2001; Winter et al. 2005, 2006; Perlut et

al. 2006; Veech 2006; Askins et al. 2007; Renfrew and Ribic 2008; Ribic et al. 2009;

Walk et al. 2010). Although the factors implicated in these declines are diverse and

complex, the primary cause can be understood as the degradation of grassland habitat

due to agricultural intensification.

Murphy (2003) demonstrated that U.S. avian population trends between 1980 and

1998 were strongly correlated with changes in agricultural land use, and more

specifically with loss of suitable habitat. This loss refers to habitat extent, but perhaps

even more so to habitat quality. Many grassland species currently rely on agricultural

lands that simulate prairie habitats, such as pastures and hayfields, which now make up

the majority of grassland habitats in the United States (Akins et al. 2007). On such

surrogate grasslands, the use of increasingly intensive farming practices (i.e. application

of chemicals, earlier and more frequent mowing of hayfields, higher stocking rates of

livestock, etc.) threaten birds that use these habitats for breeding and foraging (Murphy

2003; Askins et al. 2007). Grassland bird conservation now depends on the development

9

and adoption of progressive farming practices that balance the economic needs of farmers

and consumers with the ecological needs of the birds.

One such promising practice that has emerged over the last 30 years is rotational

grazing of livestock. Rotational grazing (RG) is a method of pasture management in

which a grazing herd is rotated through small sections of pasture, or paddocks, on a

regular basis (often daily). After grazing, each paddock is rested and left to re-grow so

that no section is overgrazed and the herd receives a constant source of nutritious

feedstock. In dairy farming, RG has been promoted as a way to optimize milk production

while protecting natural resources and improving livestock health (Aschmann and

Cropper 2007). Previous studies suggest that rotational grazing can provide multiple

environmental benefits such as reduced runoff and erosion, increased wildlife habitat and

improved animal and human health (Undersander et al. 2002). These benefits stem in

part from the fact that the perennial grass cover of pastures protects soil year round, and

requires less pesticide and fertilizer input than annual crops such as corn and soybeans

(Paine et al. 1995). Much of the literature promoting RG also refers to its potential to

benefit grassland birds while remaining profitable for the farmer (Undersander et al.

2000, 2002; Aschmann and Cropper 2007; Gillespie et al. 2009). In fact, some

Midwestern farmers have reported increases in grassland-obligate species on their farms

after replacing row crops with pastures (Jackson 2002).

Relatively few studies have directly examined the impact of rotational grazing on

bird abundance. Paine et al. (1995) compared the number of grassland bird territories

across rotationally grazed paddocks, continuously grazed pastures, and refuge areas

(groups of paddocks where grazing or mowing is deferred until after the breeding season)

on Wisconsin dairy farms. Almost twice as many territories were observed per 100 acres

in rotationally grazed paddocks versus continuously grazed pastures, and refuge areas

supported an even higher density of territories (Paine et al. 1995). Estimated nest

survival, however, was not substantially different between rotational and continuous

pastures (Paine et al. 1995), suggesting that adoption of a rotational grazing regime is

unlikely to increase bird populations over the long term.

10

Most of the remaining studies assessing the response of bird abundance to grazing

have been conducted on prairies or rangelands managed for beef production rather than

dairy pastures (Schneider 1998; Fondell and Ball 2004; Fuhlendorf et al. 2006). These

studies demonstrated that grazing may have positive or negative impacts on birds,

depending on a variety of contextual factors including management intensity, land use

history, and the structure and composition of the surrounding landscape (Best et al.

2001). Responses of bird abundance to grazing may vary by year (Schneider 1998) and

region (Winter et al. 2006), as well as by the habitat requirements of individual species,

and the specific grazing treatment used (Kantrud and Kologiski 1983). Few studies have

assessed the impact of rotational grazing and habitat variables together, and there is

currently little evidence of whether RG, in the context of a fragmented dairy landscape,

does in fact produce measureable benefits to grassland birds.

Our study used a multi-scale approach to examine bird abundance in response to

grazing intensity and habitat (i.e. land cover) variables, in order to evaluate which factors

best explained differences in relative abundance of grassland birds on dairy farms. In

particular, our primary objective was to assess whether grazing farms supported

substantially higher abundances of grassland birds than confinement farms, and whether

intensively managed RG farms in particular (as opposed to continuous grazing or other

less-intensive pasture management practices) provided a unique benefit to these species.

Understanding the value of RG to grassland birds may reveal new opportunities for bird

conservation, and may clarify the degree to which RG can enhance agricultural

sustainability in general. By using a multi-scale approach we were able to assess how the

composition of the surrounding landscape may influence the realized benefits of

rotational grazing.

METHODS

Study Sites

11

We visited 53 dairy farms from 20 May – 28 August 2009, including 20 farms in

Wisconsin (Clark County), 16 in Pennsylvania (Berks, Lancaster, Lebanon, and York

Counties), and 17 in New York (Cortland, Madison, and Tompkins Counties; Figure 1).

All counties consisted of a highly fragmented landscape with a high density of dairy

operations (Figure 2), but each region differed in its historical use of rotational grazing,

as well as its physical and cultural landscape. Clark County, in central WI, was a

relatively flat agricultural landscape, while Pennsylvania and New York were hillier and

more forested. The use of rotational grazing in the Midwest region increased steeply

during the 1990s, from approximately 7% of Wisconsin dairy farms in 1993 to 26% in

2005 (PATS 2007). The number of Northeast dairy farms using rotational grazing was

less well documented, but estimates ranged between 10-22% of operations (Knoblauch et

al. 2001; Nott 2003; Winsten et al. 2010).

Farms with " 300 milk cows were selected from a pool of willing participants in a

broader study of potential multifunctional benefits from rotational grazing (Jordan et al.

unpubl. data). In each state, farms were selected to sample a broad spectrum of

management practices, including confinement and continuous or rotational grazing. Post

data-collection, farms were grouped into three categories: confinement, low-intensity

grazing, or high-intensity grazing (Figure 3). Confinement farms were those on which

pasture provided little to none of the milking herd’s nutrition; cows were typically fed

hay or corn silage, grains, and/or total mixed rations (TMR) in the barn and, if pastured,

were not moved between different paddocks. Low-intensity grazing farms included those

on which milk cows received < 50% of their nutrition from grazing, but were largely kept

on pasture during the grazing season and were moved to a fresh pasture or paddock on a

semi-regular basis. High-intensity grazing farms were those that met three specific

criteria: 1) milk cows received ! 50% of their nutrition from pasture during the growing

season, 2) cows were rotated to a new paddock at least once a day, and 3) the farmer

spoke during structured interviews about actively managing pasture soils and/or

vegetation.

Observed management practices actually fell more along a continuum; for

example, grazing farms varied in stocking density (number of cows per acre of pasture),

12

length of grazing rotation, organic certification, and techniques used to manage soils and

vegetation on the farm. However, grouping farms into three distinct categories facilitated

useful comparisons among management styles.

Avian Species Richness and Relative Abundance

A single observer (K. Clower) surveyed farmland birds using 5-min 100m fixed-

radius point counts (Ralph et al. 1993). Survey points were allocated in ArcGIS in a

systematic random fashion, with a minimum of 6 points per farm (median = 14, max =

17). Points were placed such that the outer edge of each point circle fell within farm

property lines, and points were separated by ! 200m to avoid overlapping count circles

and thus minimize the chances of double-counting the same birds. Survey points were

located within a variety of farm habitats, including pastures, row crops, Conservation

Reserve Program (CRP) lands, exercise lots, and wetlands. However, points were not

allowed to occur on open water or interior woodland. In the field, survey points were

located using a hand-held GPS; in cases where the pre-determined point was inaccessible,

surveys were conducted from a spot as close as possible to the original point, which

occasionally resulted in small portions of the survey area falling outside of farm property

bounds. Surveys were not performed in rain or strong wind (> 6 on the Beaufort scale).

Each point was surveyed once between 0600 and 1200 local time. Wisconsin

surveys took place from 20 May to 23 June, Pennsylvania surveys from 30 June to 22

July, and New York surveys from 24 July to 28 August, concurrent with other

biophysical surveys. We recorded all birds seen or heard in the 5-min survey period,

including birds such as swallows that flew through the count circle. In order to estimate

detectability, point counts were separated into two time and two distance segments

(Ralph et al. 1993): we noted whether birds were seen or heard within the first three

minutes or the last two minutes, and whether they occurred within a 50m-radius of the

point, or between 50m and 100m. However, because detectability did not differ among

habitats (K. Clower, unpubl. data), data from both time and distance segments were

combined for subsequent analyses. Birds that flushed as the observer approached the

survey point were counted as being in the survey area. Individuals that occurred outside

13

of the point circle were tallied for use in species richness estimates, but were not used in

analyses of relative abundance. In order to assess species richness among farms, we also

recorded the presence of any additional species seen on the property at any point during

the farm visit.

Land Cover

We divided land cover into five classes: 1) pasture (continuous and rotational), 2)

row crops (e.g., corn, soybeans, buckwheat), 3) hay/small grains (grass and alfalfa

mowed for hay, or small grains; hereafter hay), 4) forest, and 5) other (open water,

wetlands, shrubland, fallow fields, developed areas such as roads and buildings, etc). We

used digitized aerial images of farms in ArcGIS to calculate the percent distribution of

each land cover class within 100m and 1200m radii of each survey point (Figure 4). In

the few cases where part of the 100m point-count circle fell outside of farm property

boundaries, land cover was expressed as the percentage of the circle within farm

property. For the majority of farm habitats, observers verified land cover during the farm

visit; where land cover was unverified, and for portions of the 1200m-radius circle that

fell outside of farm property, we used the 2009 U.S. Department of Agriculture’s

National Agriculture Statistics Service (NASS) Cropland Data Layer to assess habitat

composition.

Statistical Analysis

Species richness was estimated by counting the total number of species seen on a

farm, both during point-count surveys and throughout the course of other biophysical

surveys performed during the site visit. We used 2-way analysis of variance (ANOVA)

in program R (2.9.2) to test for significant differences in mean species richness by farm

type (confinement, low-intensity, and high-intensity) and by state.

Species abundance was a raw count of the number of individuals of each species

observed at each survey point. Abundance models were developed in SAS (9.2) for the

three most common species of grassland songbirds: Bobolink (Dolichonyx oryzivorus),

Grasshopper Sparrow (Ammodramus savannarum), and Savannah Sparrow (Passerculus

14

sandwichensis). Horned Larks (Eremophila alpestris) were observed on >10% of

surveys, but were not included in analyses as grassland obligate species because of their

unique ability to utilize cultivated fields in addition to grassland habitats (Undersander et

al. 2000). Grassland songbird species encountered at densities too low for analysis

included Eastern Meadowlark (Sturnella magna), Clay-colored Sparrow (Spizella

pallida), and Field Sparrow (Spizella pusilla).

To address our primary objective of determining whether rotational grazing farms

supported substantially higher abundances of grassland birds, we conducted 2-way

ANOVAs to examine differences in relative abundance among states and farm types

using PROC GENMOD. Count data were modeled as either Poisson or negative

binomial distributions (McDonald et al. 2000) and we included farm as a repeated effect

to account for non-independence among sampling points within each farm. We began

with a full model that included an intercept, overdispersion parameter, and effects of farm

type, state, and their interaction. We deleted overdispersion parameters and interactions

if they were non-significant (P > 0.05) or inestimable (Bobolinks and Grasshopper

Sparrows were not found in all states and so interactions were inestimable). To assess the

effect of grazing intensity on bird abundance, we estimated least-squares means from the

best supported models. Least-squares means provided an estimate of relative abundance

assuming that all three farm types had been sampled in equal proportions in all three

states.

To further explore sources of variation in grassland songbird abundance, we

modeled relative abundance as a response to Julian date and land cover at two scales

using PROC GLIMMIX. We began with full models including an intercept and variables

for proportions of crop, hay, pasture and forest at both 100m and 1200m radii, as well as

linear and quadratic terms for Julian date (Table 1). We used date rather than state

because the two variables were highly correlated (r = 0.74) and because we expected

differences in abundance to reflect seasonality more than habitat quality in a particular

state (ranges of all 3 primary species broadly overlap all 3 states included in our study

design). We simplified models by sequentially removing the least significant variable

from each model until all remaining variables were significant (P < 0.05). To illustrate

15

the effects of significant variables, we used ESTIMATE statements to predict the

abundances of each species across the entire observed range of habitat values (Figure 5).

For variables describing habitat composition, we altered one variable at a time while

forcing all habitat variables to sum to 100% and keeping unmodeled variables at their

observed relative proportions (e.g. if pasture averaged 20% and row crops averaged 40%

of the landscape, we would model pasture across the observed range of 0-85% of the

landscape, while constraining row crops to be 50% of the remainder).

RESULTS

Farm Attributes

Individual study farms varied widely in size and landscape composition (Table 2).

Farm size ranged from 46 – 1322 acres, but average farm size was approximately equal

among farm types (mean = 280, Table 2). Row crops (particularly corn) and hay

(especially alfalfa) were the predominant land cover types on confinement and low-

intensity grazing farms, together comprising on average > 60% of land on these farm

types (Table 2). High-intensity grazing farms, in contrast, had an average of > 45% of

their land in pasture and < 30% in hay and corn. This trend was particularly striking in

Wisconsin (Table 2), where the average percentage of hay on confinement and low-

intensity grazing farms was higher than in other states (mean = 42.3% and 28.7%,

respectively), and the two high-intensity grazing farms surveyed each had > 70% of their

land in pasture. The percentage of forest was low and approximately equal across farm

types (mean = 4.7%, Table 2).

Avian Species Richness and Relative Abundance

Species richness did not differ among farm types (F = 0.51, P = 0.60) or states (F

= 1.21, P = 0.31). Throughout the study period I observed 76 species of birds: 58 each in

Wisconsin and New York, and 51 in Pennsylvania. The most commonly observed species

(Table 3) included many generalists such as Rock Pigeons (Columba livia), House

16

Sparrows (Passer domesticus), Red-winged Blackbirds (Agelaius phoeniceus), and

European Starlings (Sturnus vulgaris).

Grassland songbirds were observed across all three farm types, with at least one

individual occurring on 89% of farms (Table 3). Grassland birds were observed with less

frequency and in lower abundance than other guilds, with the exception of Savannah

Sparrows, which were observed on > 80% of all farms.

Abundance of individual grassland species did not differ significantly among farm

types (Table 4). Bobolinks tended to be more common on grazing farms of either type

(low- or high-intensity) and Grasshopper Sparrows tended to be more common on high-

intensity grazing farms, but these differences were not significant (P = 0.19 and 0.12,

respectively). For each species, least-squares means indicated that bird abundances

varied between farm types by < 0.25 birds per point count circle (Table 5). Abundance of

Savannah Sparrows was significantly different among states (Table 5) and was highest in

Wisconsin, where surveys coincided with the height of the breeding season. Bobolinks

were not detected in Pennsylvania and Grasshopper Sparrows were not detected in New

York, suggesting real regional differences in abundance; however, statistical models for

these two species did not converge, which precluded calculation of significance tests.

Responses to land-cover variables differed among species, but in the majority of

models, relative abundance was positively associated with the proportion of surrogate

grassland habitats (either hayfields or pastures) and negatively associated with woody

cover (Table 6). Similarly, each species demonstrated a different response to Julian date.

Savannah Sparrow abundance declined with survey date, while Grasshopper Sparrow

abundance showed no significant relationship to date. For Bobolinks, abundances

initially declined with survey date, but spiked in late August when many species were

flocking up prior to migration (Table 6).

The proportion of hay within the count circle (Hay100) was the only variable that

was significant in all three models, and had a consistently positive effect on bird

abundances (Table 6). Percentage pasture within the count circle was positively

associated with abundance of Savannah Sparrows and Grasshopper Sparrows, but was

not a significant predictor of Bobolink abundance. Savannah Sparrow abundance also

17

showed a negative response to Crop100 and to Forest100 (Figure 5). Abundance of

Bobolinks and Grasshopper Sparrows was not significantly affected by Crop100 or

Forest100 (Table 6).

Savannah Sparrows and Grasshopper Sparrows showed a significant response to

at least one variable at the 1200m scale (Table 6). Hay1200 was positively associated

with abundance of Savannah Sparrows, and pasture1200 had a strong positive effect on

Grasshopper Sparrows. Abundance of Savannah Sparrows was positively associated

with Crop1200. As with the 100m scale, Bobolinks and Grasshopper Sparrows showed

no significant response to Forest1200 (Table 6).

DISCUSSION

Impacts of Grazing Intensity and Land Cover

The primary objective of this study was to assess the impact of rotational grazing

on grassland songbirds. We found no evidence that farms using rotational grazing (i.e.

high-intensity grazing farms) supported meaningfully higher abundances of grassland

birds than occurred on other farms. Although our data suggest a slight preference of

grassland birds for grazing farms (either low- or high-intensity) over confinement farms,

this effect was not significant for any individual species. In fact, the observed differences

in mean bird abundance between grazing and confinement farms were on average less

than 0.31 birds/ha (Table 4). Our models suggest that these differences are related more

to land cover (i.e. amount of pasture and hayland) than to management practices per se.

Our findings are consistent with previous studies of rotational grazing. In a 2-

year study in North Dakota, Schneider (1998) compared relative abundance of grassland

passerines between pastures using a rotational grazing regime and those using traditional

season-long grazing. For the ten species analyzed, relative abundance did not differ

significantly between grazing regimes in either year. However, a sub-set of “grazing

sensitive” species, including Savannah Sparrows, Grasshopper Sparrows and Bobolinks,

18

was significantly more abundant on RG pastures in one year (Schneider 1998).

Schneider concluded that some of the observed variation in bird abundance was related to

vegetation structure, which varied with annual precipitation. A more recent study by

Bleho (2009), at Grassland National Park in Saskatchewan, Canada, found that grazing

both positively and negatively influenced abundance of grassland birds, depending on

species and grazing treatment. Bleho (2009) also found evidence that grazing indirectly

influenced birds by altering habitat heterogeneity and structure. Both of these studies

suggest that vegetation structure and other habitat variables may be better predictors of

bird abundance than grazing intensity, although the two factors are interrelated.

In our study, habitats and their spatial distributions appeared to have a stronger

impact than grazing intensity on bird abundance. Abundance was positively associated

with increases in the proportions of hay and/or pasture within a 100m radius, and

negatively associated with increases in the amount of forest within 100m (Figure 5). Our

estimates predicted that increasing the percentage of hay or pasture from 0 – 100% within

a point-count circle would increase mean bird abundance by at least 80%, from 0.43 –

0.92 birds per survey in the case of Savannah Sparrows (Figure 5). However, Savannah

Sparrow abundance decreased steeply with increases in forest within 100m and was

predicted to be essentially zero at proportions of forest above 0.6 (Figure 5). These

findings were expected considering that all three species modeled commonly breed and

forage in hayfields and pastures, whereas forested areas are considered unsuitable habitat

(Paine et al. 1996; Sample and Mossman 1997). We also observed a weak negative

association between bird abundance and the proportion of crop within 100m, although

both Savannah Sparrows and Bobolinks were occasionally observed foraging in crop

fields such as corn or soybeans.

Two out of three of our models included variables at both the 100m and 1200m

scales. One of the research questions that we were able to address by using a multi-scale

model was the importance of landscape context in determining local abundance. On a

practical level, we hypothesized that even if grazing farms provided a measureable

benefit to grassland birds, that potential benefit might not be realized until the number of

grazing farms (and thus the total area of pasture) within a landscape reached a certain

19

threshold. For Grasshopper Sparrows, abundance showed a weak response to variables at

the 100m scale, but demonstrated a strong positive association with the proportion of

pasture within a 1200m radius.

Our findings are consistent with other studies that have demonstrated the

importance of a multi-scale approach. Many researchers (e.g. Mazerolle and Villard

1999; Best et al. 2001; Davis 2004; Cunningham and Johnson 2006; Renfrew and Ribic

2008) have found that birds respond not only to microhabitat variables such as vegetation

type and density within 100m, but also to macrohabitat and landscape variables such as

patch size and shape, cover type diversity, and habitat fragmentation at up to 1600m from

a survey point. Ribic and Sample (2001) found that Grasshopper Sparrows and Savannah

Sparrows responded to a combination of field and landscape variables, while Bobolinks

responded primarily to landscape-scale variables. In contrast, Bobolink abundance in our

study was not associated with any land cover variables at the 1200m scale.

Because of the correlation between land cover classes and farm types in our study,

we were unable to directly address interactions between management activities and

habitat features. To fully understand the impact of rotational grazing on grassland birds,

additional research is needed to explore the effects of various pasture management

strategies on habitat features such as vegetative structure and composition, forage and

nest-site availability, and the level of fragmentation in the landscape.

Limitations

Several study limitations deserve mention. First of all, our study was designed to

assess multiple biophysical and social factors across a broad sample of farms. Due to

logistical constraints, we were only able to visit each farm once. Bird abundance changes

both annually and seasonally, responding to weather patterns, changes in breeding habitat

quality and availability, and conditions on wintering grounds (Ahlering et al. 2009; Butler

et al. 2009). In addition, birds become less detectable later in the season as singing

behavior by territorial males diminishes (Ralph et al. 1993). Conducting point-count

surveys only once at each point location is unlikely to be sufficient for capturing long-

20

term trends, particularly at New York and Pennsylvania sites visited late in the breeding

season after territorial singing behavior was beginning to wane.

Second, interpretation of our results is based on the assumption that bird

abundance is a surrogate measure of habitat quality, when in fact there are situations in

which this assumption does not hold true (Van Horne 1983). For example, Bollinger et

al. (1990) demonstrated that hayfields could be a sink habitat for Bobolinks, as mowing

of hayfields has contributed to the decline in populations of Bobolinks, both directly

through mortality caused during mowing and indirectly through increased nest

abandonment and predation after mowing. In another study, mowing caused 99% of

active Savannah Sparrow nests to fail, and early haying was linked to reduced

reproductive success (Perlut 2007). An observer performing surveys the day before

mowing might find very different results than if s/he had visited the day after mowing.

Finally, reversing the decline of grassland bird populations will require accurate

information on the long-term impact of grazing on bird population demographics (i.e.

survival and reproductive success), in addition to measures of abundance. Our study

examined broad trends in bird abundance across farm types, but did not address

population demographics in relation to farm management. For dairy farms to contribute

to the recovery of grassland birds, they must not only attract breeding birds, but also

provide productive breeding habitat.

Despite these limitations, our results suggest that differences in abundance of

grassland birds among farm types were modest at best, and that management practices

must be understood in the context of the surrounding landscape.

MANAGEMENT IMPLICATIONS

Rotational grazing farms had a majority of their land devoted to pasture, whereas

confinement and traditional grazing farms had more land in corn and hay. Although this

large amount of pasture on rotationally-grazed farms was attractive to grassland birds, it

was compensated to some extent on other farm types by increased acreage of hay, which

21

was also attractive to grassland birds. Hence, we saw little effect of management

practices on grassland bird abundance at the farm scale.

Our models demonstrate that grassland birds responded to variables at multiple

scales, suggesting that management must also take place at multiple scales. For example,

Grasshopper Sparrow abundance was negligible when the proportion of pasture within a

1200m radius was < 0.4, but when the proportion of pasture reached 0.5 or higher,

Grasshopper Sparrow abundance increased dramatically (Figure 5). This finding implies

that management changes at the field scale alone will not have a meaningful impact on

Grasshopper Sparrow populations, and that whole farm- and landscape-scale planning is

needed (Sample et al. 2003).

Our data show that grassland birds can benefit from increases in the total area of

surrogate grassland, such as pasture and hayfields, on the landscape. Grazing farms,

almost by definition, can provide this type of habitat. However, grazing may not be a

viable or desirable option for all dairy farmers. Furthermore, the benefits provided by

increasing grassland habitats on a small number of farms are unlikely to have a

substantial impact on bird populations, particularly in the context of a highly fragmented

agricultural landscape. Future research for conservation planning and policy

development should focus on the landscape-scale to ensure that conservation actions are

most effective.

22

LITERATURE CITED

Agricultural Technology and Family Farm Institute. 1996. Management intensive rotational grazing in Wisconsin: Who’s doing it? University of Wisconsin-Madison and University of Wisconsin – Cooperative Extension, Wisconsin Family Farm Facts, No. 5.

Ahlering, M. A., D. H. Johnson and J. Faaborg. 2009. Factors associated with arrival

densities of Grasshopper sparrow (Ammodramus savannarum) and Baird’s sparrow (A. bairdii) in the upper Great Plains. Auk, 126: 799-808.

Aschmann, S. and J. B. Cropper, eds. 2007. Profitable grazing-based dairy systems. U.S.

Department of Agriculture, Natural Resources Conservation Service, Range and Pasture technical note 1.

Askins, R. A., F. Chavez-Ramirez, B. C. Dale, C. A. Haas, J. R. Herkert, F. L. Knopf and

P. D. Vickery. 2007. Conservation of grassland birds in North America: understanding ecological processes in different regions. Ornithological Monographs, no. 64.

Battin, J. and J. J. Lawler. 2006. Cross-scale correlations and the design and analysis of

avian habitat selection studies. Condor, 108: 59-70. Best, L. B., T. M. Bergin and K. E. Freemark. 2001. Influence of landscape composition

on bird use of rowcrop fields. Journal of Wildlife Management, 65: 442-449. Bignal, E. M. and D. I. McCracken. 1996. Low-intensity farming systems in the

conservation of the countryside. Journal of Applied Ecology 33: 413-424. Bills, N. and D. Gross. 2005. Sustaining multifunctional agricultural landscapes:

comparing stakeholder perspectives in New York (US) and England (UK). Land Use Policy, 22: 313-321.

Bleho, B. 2009. Passerine relationships with habitat heterogeneity and grazing at multiple

scales in northern Mixed-grass prairie. Master’s thesis, University of Manitoba, Winnipeg, Manitoba.

Bollinger, E. K., P. B. Bollinger and T. A. Gavin. 1990. Effects of hay-cropping on

eastern populations of the Bobolink. Wildlife Society Bulletin, 18:142-150.

23

Boody, G., B. Vondracek, D. A. Andow, J. Zimmerman, M. Krinke, J. Westra, and P. Welle. 2005. Multifunctional agriculture in the United States. Bioscience 55:27-38.

Brennan, L. A. and W. P. Kuvlesky, Jr. 2005. Invited Paper: North American grassland

birds: an unfolding conservation crisis? Journal of Wildlife Management, 69(1): 1-13.

Butler, A. B., J. A. Martin, W. E. Palmer and J. P. Carroll. 2009. Winter use of south

Florida dry prairie by two declining grassland passerines. The Condor, 111(3): 511-522.

Cunningham, M. A. and D. T. Johnson. 2006. Proximate and landscape factors influence

grassland bird distributions. Ecological Applications, 16(3): 1062-1075. 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(4):1130-1145.

Fondell, T. F. and I. J. Ball. 2004. Density and success of bird nests relative to grazing on

western Montana grasslands. Biological Conservation, 117: 203-213. Fuhlendorf , S. D., W. C. Harrell, D. M. Engle, R. G. Hamilton, C. A. Davis and D. M.

Leslie Jr. 2006. Should heterogeneity be the basis for conservation? Grassland bird response to fire and grazing. Ecological Applications, 16(5): 1706-1716.

Gillespie, J., R. Nehring, C. Hallahan and C. Sandretto. 2009. Pasture-based dairy

systems: Who are the producers and are their operations more profitable than conventional dairies? Journal of Agricultural and Resource Economics, 34(3): 412-427.

Helzer, C. J. and D. E. Jelinski. 1999. The relative importance of patch area and

perimeter-area ratio to grassland breeding birds. Ecological Applications, 9(4): 1448-1458.

Herkert, J. R. 1995. An analysis of Midwestern breeding bird population trends: 1966-

1993. American Midland Naturalist, 134(1): 41-50. Herkert, J. R. 1997. Bobolink Dolichonyx oryzivorus population decline in agricultural

landscapes in the Midwestern USA. Biological Conservation, 80: 107-112. Jackson, D. L. 2002. The farm as natural habitat. Pages 13-26 in Jackson, D. L. and L. L.

Jackson, eds. The Farm as Natural Habitat. Island Press, Washington D. C., USA.

24

Jobin, B., J-L. DesGranges and C. Boutin. 1998. Farmland habitat use by breeding birds

in Southern Quebec. Canadian Field-Naturalist, 112(4):611-618. Jordan, N. and K. D. Warner. 2010. Enhancing the multifunctionality of US agriculture.

Bioscience, 60(1): 60-66. Kantrud, H. A. and R. L. Kologiski. 1983. Avian associations of the northern Great

Plains grasslands. Journal of Biogeography, 10: 331-350. Kriegl, T. and R. McNair. 2005. Pastures of plenty: Financial performance of Wisconsin

grazing dairy farms. Center for Integrated Agricultural Systems, Center for Dairy Profitability, and Program on Agricultural Technology Studies, University of Wisconsin-Madison.

Liu, J., T. Dietz, S. R. Carpenter, M. Alberti, C. Folke, E. Moran, A. N. Pell, P.

Deadman, T. Kratz, J. Lubchenco, E. Ostrom, Z. Ouyang, W. Provencher, C. L. Redman, S. H. Schneider and W. W. Taylor. 2007. Complexity of coupled human and natural systems. Science, 317: 1513-1516.

Lyons, J., B. M. Weigel, L. K. Paine and D. J. Undersander. 2000. Influence of intensive

rotational grazing on bank erosion, fish habitat quality, and fish communities in southwestern Wisconsin trout streams. Journal of Soil and Water Conservation 55(3): 271.

Mazerolle, M. J. and M-A. Villard. 1999. Patch characteristics and landscape context as

predictors of species presence and abundance: a review. Ecoscience, 6(1):117-124.

Morse, C. E. 2010. Strategies for promoting working landscapes in North America and

Europe: A report for the Vermont Council on Rural Development. Available: http://vtrural.org/programs/policy-councils/working-landscape. Accessed 3 March 2011.

Murphy, M. 2003. Avian population trends within the evolving agricultural landscape of

eastern and central United States. The Auk 120: 20. Naveh, Z. 1994. Interactions of landscapes and cultures. Landscape and Urban Planning,

32: 43-54. Nott, S. B. 2003. Evolution of dairy grazing in the 1990’s. Staff Paper #2003-07.

Department of Agricultural Economics, Michigan State University.

25

Paine, L., G. A. Bartelt, D. J. Undersander and S. A. Temple. 1995. Agricultural Practices for the Birds. Passenger Pigeon, 57: 77-87.

Paine, L., D. J. Undersander, D. W. Sample, G. A. Bartelt and T. A. Schatteman. 1996.

Cattle trampling of simulated ground nests in rotationally grazed pastures. Journal of Rangeland Management, 49: 294-300.

Program on Agricultural Technology Studies. 2007. Structural change in WI dairy 1987-

2007: divergence in size and system. Fact Sheet No. 25. Perlut, N. G., A. M. Strong, T. M. Donovan and N. J. Buckley. 2006. Grassland

songbirds in a dynamic management landscape: behavioral responses and management strategies. Ecological Applications, 16(6):2235-2247.

Perlut, N. G. 2007. Effects of hayfield management on grassland songbirds: behavioral

responses and population processes. Dissertation, University of Vermont. Peterjohn, B. G. and J. R. Sauer. 1999. Population status of North American grassland

birds from the North American Breeding Bird Survey, 1966-1996. Studies in Avian Biology, 19: 27-44.

Ralph, C. J., G. R. Geupel, P. Pyle, T. E. Martin, and D. F. DeSante. 1993. Handbook of

field methods for monitoring landbirds. General Technical Report PSW-GTR-144. Albany, CA: Pacific Southwest Research Station, Forest Service, U.S. Department of Agriculture.

Renfrew, R. B. and C. I. Ribic. 2008. Multi-scale models of grassland passerine

abundance in a fragmented system in Wisconsin. Landscape Ecology 23(2): 181-193.

Renting, H., W. A. H. Rossing, J. C. J. Groot, J. D. Van der Ploeg, C. Laurent, D.

Perraud, D. J. Stobbelaar and M. K. Van Ittersum. 2009. Exploring multifunctional agriculture: a review of conceptual approaches and prospects for an integrative transitional framework. Journal of Environmental Management, 90: 5112-5123.

Ribic, C. A. and D. W. Sample. 2001. Associations of grassland birds with landscape

factors in Southern Wisconsin. American Midland Naturalist, 146:105-121. Ribic, C. A., R. R. Koford, J. R. Herkert, D. H. Johnson, N. D. Niemuth, D. E. Naugle, K.

K. Bakker, D. W. Sample and R. B. Renfrew. 2009. Area sensitivity in North American grassland birds: patterns and processes. The Auk, 126(2):233-244.

Sample, D. W., and M. J. Mossman. 1997. Managing habitat for grassland birds - a guide

26

for Wisconsin. Wisconsin Department of Natural Resources, Madison, WI, PUBL-SS-925-97. 154 pp. Jamestown, ND: Northern Prairie Wildlife Research Center Online. http://www.npwrc.usgs.gov/resource/birds/wiscbird/index.htp (Version 03JUN2002)

Sample, D. W., C. A. Ribic and R. B. Renfrew. 2003. Linking landscape management

with the conservation of grassland birds in Wisconsin. Pg. 359-385 in Landscape Ecology and Resource Management, J. A. Bissonette and I. Storch, eds., Island Press, Washington, DC, 2003.

Schneider, N. A. 1998. Passerine use of grasslands managed with two grazing regimes on

the Missouri Coteau in North Dakota. Master’s thesis, South Dakota State University.

Selman, P. and M. Knight. 2006. On the nature of virtuous change in cultural landscapes:

exploring sustainability through qualitative models. Landscape Research, 31(3): 295-307.

Sovell, L., B. Vondracek, J. A. Frost and K. G. Mumford. 2000. Impacts of rotational

grazing and riparian buffers on physicochemical and biological characteristics of Southeastern Minnesota, USA, streams. Environmental Management, 26(6): 629-641.

Taylor, J. and J. Foltz. 2006. Grazing in the dairy state: Pasture use in the Wisconsin

dairy industry, 1993-2003. Center for Integrated Agricultural Systems and Program on Agricultural Technology Studies, University of Wisconsin-Madison.

Undersander, D., B. Albert, D. Cosgrove, D. Johnson and P. Peterson. 2002. Pastures for

profit: a guide to rotational grazing. Cooperative Extension Publishing, University of Wisconsin-Extension.

Van Horne, B. 1983. Density as a misleading indicator of habitat quality. The Journal of

Wildlife Management, 47(4): 893-901. Veech, J. A. 2006. A comparison of landscapes occupied by increasing and decreasing

populations of grassland birds. Conservation Biology, 20(5): 1422-1432. Vejre, H., J. Abildtrup, E. Andersen, P. Andersen, J. Brandt, A. Busck, T. Dalgaard, B.

Hasler, H. Huusom, L. S. Kristensen, S. P. Kristensen and S. Praestholm. 2007. Multifunctional agriculture and multifunctional landscapes – land use as an interface. Pgs. 93-104 in Mander, U., H. Wiggering and K. Helming (eds), Multifunctional Land Use.

27

Vickery, P. D. and J. R. Herkert. 2001. Recent advances in grassland bird research: where do we go from here? The Auk, 118(1):11-15.

Walk, J. W. and R. E. Warner. 2000. Grassland management for the conservation of

songbirds in the Midwestern USA. Biological Conservation 94(2): 165-172. Walk, J. W., E. L. Kershner, T. J. Benson and R. E. Warner. 2010. Nesting success of

grassland birds in small patches in an agricultural landscape. The Auk, 127(2): 328-334.

Winsten, J. R., C. D. Kerchner, A. Richardson, A. Lichau and J. M. Hyman. 2010. Trends

in the Northeast dairy industry: large-scale modern confinement feeding and management-intensive grazing. Journal of Dairy Science, 93:1759-1769.

Winter, M., D. H. Johnson and J. A. Shaffer. 2005. Variability in vegetation effects on

density and nesting success of grassland birds. Journal of Wildlife Management, 69(1): 185-197.

Winter, M., D. H. Johnson, J. A. Shaffer, T. M. Donovan and W. D. Svedarsky. 2006.

Patch size and landscape effects on density and nesting success of grassland birds. The Journal of Wildlife Management, 70(1): 158-172.

United States Department of Agriculture, Census of Agriculture. Farm Acreage: 1900-

1997. Available: http://www.nass.usda.gov/Publications/Trends_in_U.S._ Agriculture/Farm_Numbers/index.asp. Accessed 4 March 2011.

28

Table 1. Description, mean and range of values for explanatory variables used to model grassland bird abundance.

Variable Description Mean Range

Julian Julian date of survey (140 = 20 May, 240 = 28 Aug) 187.73 140 - 240

Intensity Farm type/Level of grazing intensity where Confinement = 1, Low-intensity grazing = 2, High-intensity grazing = 3

1.67 1 - 3

Hay100 Proportion of hay (mowed grass or alfalfa, or small grains) within 100m-radius of survey point. 0.31 0 - 1

Pasture100 Proportion of pasture (continuous and rotational) within 100m-radius of survey point. 0.25 0 - 1

Forest100 Proportion of forest within 100m-radius of survey point. 0.05 0 - 0.95

Crop100 Proportion of annual row crops within 100m-radius of survey point. 0.31 0 - 1

Hay1200 Proportion of hay within 1200m-radius of survey point. 0.25 0.002 - 0.49

Pasture1200 Proportion of pasture within 1200m-radius of survey point. 0.18 0 - 0.85

Forest1200 Proportion of forest within 1200m-radius of survey point. 0.22 0.002 - 0.78

Crop1200 Proportion of crop within 1200m-radius of survey point. 0.23 0.001 - 0.52

29

Table 2. Acreage and land cover distribution on study farms, by farm type and by state.

Total acreage % Hay % Pasture % Crop % Forest % Other Farm Type Mean Range Mean Range Mean Range Mean Range Mean Range Mean Range

Confinement (n = 31) 298 46-1322 33 4-64 9 0-35 33 5-67 15 0-41 9 1-25

Low-intensity grazing (n = 7)

225 120-336 33 14-56 21 17-25 29 0-62 8 0-24 9 3-24

High-intensity grazing (n = 15)

273 50-676 17 0-68 47 12-96 8 0-28 18 0-45 10 4-20

State

Wisconsin (n = 20) 297 80 – 1322 34 0 – 59 17 0 – 94 30 0 – 62 11 0 – 31 8 1 - 24

Pennsylvania (n = 16) 138 46 - 376 20 0 – 45 34 5 – 96 31 0 – 67 7 0 – 27 8 2 - 25

New York (n = 17) 397 121 - 1232 30 12 – 68 15 0 – 38 15 0 – 45 27 7 – 45 12 5 - 25

30

Table 3. Bird species observed on at least 14 of the 53 study farms. * Denotes grassland-obligate species.

Bird species # farms occurred on %

Barn swallow (Hirundo rustica) 48 91 American robin (Turdus migratorius) 45 85 American goldfinch (Spinus tristis) 44 83 Savannah sparrow (Passerculus sandwichensis) * 44 83 Song sparrow (Melospiza melodica) 44 83 Rock pigeon (Columba livia) 43 81 Red-winged blackbird (Agelaius phoeniceus) 40 75 House sparrow (Passer domesticus) 37 70 European starling (Sturnus vulgaris) 35 66 Gray catbird (Dumetella carolinensis) 35 66 Mourning dove (Zenaida macrocoura) 35 66 Brown-headed cowbird (Molothrus ater) 34 64 American crow (Corcus brachyrhynchos) 32 60 Killdeer (Charadrius vociferus) 32 60 Blue jay (Cyanocitta cristata) 29 55 Eastern kingbird (Tyrannus tyrannus) 27 51 Chipping sparrow (Spizella passerina) 24 45 Red-tailed hawk (Buteo jamaicensis) 24 45 Tree swallow (Tachycineta bicolor) 23 43 Horned lark (Eremophila alpestris) 22 42 Eastern wood-pewee (Contopus virens) 19 36 Bobolink (Dolichonyx oryzivorus) * 18 34 Cliff swallow (Petrochelidon pyrrhonata) 17 32 Eastern phoebe (Sayornis phoebe) 17 32 Indigo bunting (Passerina cyanea) 17 32 Northern cardinal (Cardinalis cardinalis) 17 32 Turkey vulture (Cathartes aura) 16 30 Eastern meadowlark (Sturnella magna) * 16 30 House wren (Troglodytes aedon) 15 28 Grasshopper sparrow (Ammodramus savannarum) * 14 26

31

Table 4. Mean observed abundance per point-count circle (3.14 ha) of grassland bird species most commonly observed on study farms. Includes abundances from all survey points regardless of land-cover class present.

Observed abundance (# individuals / point-count circle)

Farm Type Savannah Sparrow Horned Lark

Grasshopper Sparrow

Bobolink Eastern Meadowlark

Confinement (n = 396) 1.00 ± 0.07 0.26 ± 0.04 0.06 ± 0.01 0.05 ± 0.26 0.01 ± 0.003

Low-intensity grazing (n = 98) 1.77 ± 0.16 0.30 ± 0.06 0.02 ± 0.01 0.19 ± 0.04 0.07 ± 0.03

High-intensity grazing (n = 175) 0.69 ± 0.09 0.05 ± 0.02 0.26 ±0.05 0.15 ± 0.05 0.06 ± 0.02

32

Table 5. Least-squares means and 95% confidence intervals of relative abundance by farm type and state. Significance tests are from a 2-way ANOVA.

Bird Species Savannah Sparrow Bobolink Grasshopper Sparrow

Predictor variable LS Mean 95% CI LS Mean 95% CI LS Mean 95% CI

Confinement 0.47 0.35 - 0.62 0.05 0.03 - 0.11 0.06 0.02 - 0.13 Low-intensity grazing 0.46 0.30 - 0.69 0.19 0.09 - 0.44 0.02 0.00 - 0.12 High-intensity grazing 0.68 0.44 - 1.04 0.15 0.05 - 0.50 0.26 0.12 - 0.57 Farm type effect (!2, P) 3.03 0.22 3.35 0.19 4.32 0.12 WI 2.30 1.91 - 2.78 0.20 0.11 - 0.37 0.01 0.00 - 0.04 PA 0.26 0.13 - 0.52 0 0 0.36 0.21 - 0.63 NY 0.24 0.16 - 0.36 0.06 0.02 - 0.14 0 0 State effect* (!2, P) 23.41 0.0001 Inestimable Inestimable

* State effects were inestimable for two species due to missing values in one state (models did not converge).

33

Table 6. Best supported models of relative abundance of grassland birds in relation to land-cover variables at 100m and 1200m radii.

Species Variable ! SE p Savannah Sparrow Intercept 1.50 1.12 0.1886

Julian -0.02 0.00 < 0.0001 Crop1 0.75 0.34 0.0279 Hay1 1.26 0.34 0.0002 Pasture1 1.41 0.36 < 0.0001 Forest1 -2.83 0.87 0.0012 Crop12 1.93 0.96 0.0452 Hay12 2.82 0.78 0.0003

Bobolink Intercept 40.53 12.71 0.0025 Julian -0.46 0.14 0.0008 Julian*Julian 0.00 0.0004 0.0013 Hay1 1.62 0.43 0.0002

Grasshopper Sparrow Intercept -5.91 0.65 < 0.0001 Hay1 1.40 0.64 0.0292 Pasture1 1.19 0.51 0.0200 Pasture12 6.68 1.14 < 0.0001

34

Figure 1. Locations of counties (shaded) containing study sites in Wisconsin, Pennsylvania and New York.

WI NY

PA

35

Figure 2. Landscape views of Clark County, WI (A), Lancaster County, PA (B) and Madison County, NY (C). A. B. C.

36

Figure 3. Percent distribution of farm types surveyed in Wisconsin, Pennsylvania and New York.

37

Figure 4. Aerial photo of a study farm, showing bird survey points (yellow) with 100m (red) and 1200m (blue) radii. Farm fields are outlined in orange.

38

Figure 5. Least-squares means and standard errors of bird abundance in response to significant habitat variables, for Savannah Sparrow (A), Bobolink (B) and Grasshopper Sparrow (C). A.

39

B. C.