<%BANNER%>

Are Barn Owls (Tyto alba) Biological Controllers of Rodents in the Everglades Agricultural Area?

Permanent Link: http://ufdc.ufl.edu/UFE0022175/00001

Material Information

Title: Are Barn Owls (Tyto alba) Biological Controllers of Rodents in the Everglades Agricultural Area?
Physical Description: 1 online resource (77 p.)
Language: english
Creator: Martin, Jason
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2009

Subjects

Subjects / Keywords: agriculture, alba, barn, biological, box, breeding, control, nest, owl, rodent, success, sugarcane, tyto
Wildlife Ecology and Conservation -- Dissertations, Academic -- UF
Genre: Wildlife Ecology and Conservation thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Biological control is defined as the use of living organisms to reduce abundance of pest species. The ability of vertebrate predators to control vertebrate pests is limited, however the potential of vertebrates to reduce prey abundance is maximized when predators are capable of reproducing rapidly as prey abundance increases, when their movement is not restricted by territoriality, and if they have high foraging efficiency. Barn Owls may be effective biological controllers of rodents because they exhibit these traits. Nest boxes have been installed throughout the Everglades Agricultural Area (EAA) in south Florida to increase Barn Owl abundance with the intent of reducing pest rodents in sugarcane fields. This program has successful increased the regional abundance of owls, but it is unknown if areas with more owls have lower rodent abundances than areas without owls. Nest boxes are useful tools for expanding nest site availability for many species of birds. However, birds may be drawn to low quality sites by poorly placed boxes. Both immediate quality and long-term stability of habitat must be considered when locating nest boxes or they may act as ecological traps, resulting in low reproductive fitness for the population. When most Barn Owls in the EAA select nesting locations, harvesting activities have not begun and rodents, the primary prey of the owls, are abundant in crop fields. However, harvest occurs adjacent to most active nest boxes before nesting is completed and therefore most nests are exposed to a rapid decline in food availability. This may have negative consequences to Barn Owls in the form of lowered nest success. This study indicates that Barn Owls are not capable of reducing rodent abundance. Reproductive capacity of rodents far outpaces removal of rodents via predation. Furthermore, nest survivorship for Barn Owls was lower following sugarcane harvest when compared pre-harvest conditions. Long-term data is needed to test for the actual occurrence of an ecological trap, however this study demonstrates that a mechanism for a trap exists. If crop harvest does not result in lowered fitness of Barn Owl populations, then nest boxes may be beneficial conservation tools for Barn Owls.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Jason Martin.
Thesis: Thesis (Ph.D.)--University of Florida, 2009.
Local: Adviser: Branch, Lyn C.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2011-05-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2009
System ID: UFE0022175:00001

Permanent Link: http://ufdc.ufl.edu/UFE0022175/00001

Material Information

Title: Are Barn Owls (Tyto alba) Biological Controllers of Rodents in the Everglades Agricultural Area?
Physical Description: 1 online resource (77 p.)
Language: english
Creator: Martin, Jason
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2009

Subjects

Subjects / Keywords: agriculture, alba, barn, biological, box, breeding, control, nest, owl, rodent, success, sugarcane, tyto
Wildlife Ecology and Conservation -- Dissertations, Academic -- UF
Genre: Wildlife Ecology and Conservation thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Biological control is defined as the use of living organisms to reduce abundance of pest species. The ability of vertebrate predators to control vertebrate pests is limited, however the potential of vertebrates to reduce prey abundance is maximized when predators are capable of reproducing rapidly as prey abundance increases, when their movement is not restricted by territoriality, and if they have high foraging efficiency. Barn Owls may be effective biological controllers of rodents because they exhibit these traits. Nest boxes have been installed throughout the Everglades Agricultural Area (EAA) in south Florida to increase Barn Owl abundance with the intent of reducing pest rodents in sugarcane fields. This program has successful increased the regional abundance of owls, but it is unknown if areas with more owls have lower rodent abundances than areas without owls. Nest boxes are useful tools for expanding nest site availability for many species of birds. However, birds may be drawn to low quality sites by poorly placed boxes. Both immediate quality and long-term stability of habitat must be considered when locating nest boxes or they may act as ecological traps, resulting in low reproductive fitness for the population. When most Barn Owls in the EAA select nesting locations, harvesting activities have not begun and rodents, the primary prey of the owls, are abundant in crop fields. However, harvest occurs adjacent to most active nest boxes before nesting is completed and therefore most nests are exposed to a rapid decline in food availability. This may have negative consequences to Barn Owls in the form of lowered nest success. This study indicates that Barn Owls are not capable of reducing rodent abundance. Reproductive capacity of rodents far outpaces removal of rodents via predation. Furthermore, nest survivorship for Barn Owls was lower following sugarcane harvest when compared pre-harvest conditions. Long-term data is needed to test for the actual occurrence of an ecological trap, however this study demonstrates that a mechanism for a trap exists. If crop harvest does not result in lowered fitness of Barn Owl populations, then nest boxes may be beneficial conservation tools for Barn Owls.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Jason Martin.
Thesis: Thesis (Ph.D.)--University of Florida, 2009.
Local: Adviser: Branch, Lyn C.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2011-05-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2009
System ID: UFE0022175:00001


This item has the following downloads:


Full Text

PAGE 1

1 ARE BARN OWLS ( Tyto alba ) BIOLOGICAL CONTROLLERS OF RODENTS IN THE EVERGLADES AGRICULTURAL AREA? By JASON MICHAEL MARTIN A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2009

PAGE 2

2 2009 Jason Michael Martin

PAGE 3

3 To Suzanne – I wouldn’t be the person that I am today without you. Thank you!

PAGE 4

4 ACKNOWLEDGMENTS The Florida Sugarcane League, the Institute of Food and Agricultural Sciences – University of Florida, the Department of W ildlife Ecology and Conservation – University of Florida, the Florida Ornithological Society, and the Florida Wildlife Federation provided financial support for this research. Invaluable field help was provided by Jay Parker, Colleen O’Donnell, Joseph Sampson, and Lucy Manning. I am indebted to Jim Shine, Wayne Boynton, Modesto Ulloa, Jorge Bastanzuri, Luis Girado, Luis Entrialgo, Frank Dowdle, Gina Zimmerman, the Sugarcane Growers Cooperative of Florida, Fl orida Crystals Corp., U.S. Sugar Corp., and the faculty and staff of the Everglades Research and Education Center for logistical support. Thanks to my committee members Dr. Lyn Branch, Dr. Ri chard Raid, Dr. Katie Sieving, Dr. Ken Meyer, and Dr. Madan Oli. Research protocol D427 was approved by the UF Institutional Animal Care and Use Committee. Bird banding was conducted under permit # 22689 USGS Patuxent Wildlife Research Center Bird Banding Laboratory. Finally, my thanks to Suzanne Beyeler for all of her support and guidance.

PAGE 5

5 TABLE OF CONTENTS page ACKNOWLEDGMENTS...............................................................................................................4 LIST OF TABLES................................................................................................................. ..........6 LIST OF FIGURES................................................................................................................ .........7 CHAPTER 1 INTRODUCTION............................................................................................................ .10 2 QUANTIFYING DAMAGE POTENTIAL OF THREE RODENT SPECIES ON SUGARCANE...................................................................................................................14 Introduction.................................................................................................................. ......14 Methods....................................................................................................................... .......15 Results....................................................................................................................... .........16 Discussion.................................................................................................................... ......17 3 AN ASSESSMENT OF BARN OWLS AS BIOLOGICAL CONTROLLERS OF PEST RODENTS...............................................................................................................23 Introduction.................................................................................................................. ......23 Research Design............................................................................................................... ..26 Methods....................................................................................................................... .......27 Study Location and Species.....................................................................................27 Study Design........................................................................................................ ....28 Data Analysis....................................................................................................... ....32 Results....................................................................................................................... .........35 Discussion.................................................................................................................... ......36 4 NEST BOXES, BARN OWLS, AND AGRICULTURE: DOES CROP HARVEST LEAD TO LOWERED REPRODUCTIVE SUCCESS?..................................................47 Introduction.................................................................................................................. ......47 Research Design............................................................................................................... ..49 Methods....................................................................................................................... .......50 Results....................................................................................................................... .........55 Discussion.................................................................................................................... ......56 5 CONCLUSION..................................................................................................................6 6 LIST OF REFERENCES............................................................................................................. ..69 BIOGRAPHICAL SKETCH.........................................................................................................77

PAGE 6

6 LIST OF TABLES Table page 2-1 Mass of rodents test ed in feeding trials..............................................................................20 2-2 ANOVA examining the influence of species, sex, mass, and mean air temperature on grams of sugarcane consumed per gram of body mass by rats during feeding trials........20 3-1 Life history parameters for cotton rats derived from Montague et al. 1990 and Whitaker and Hamilton 1998, and modified to fit abundance data collected in fields that did not have Barn Owls...............................................................................................40 3-2 Competing Huggins closed-capture models ( AICc < 2) for cotton rats, roof rats, and rice rats during early, mid, a nd late sugarcane growth phases....................................41 3-3 Proc Mixed repeated measures ANOVA examining the influence of Barn Owl density and sugarcane growth on abunda nce of three species of rodents..........................42 3-4 Pair-wise least square mean comparisons among study site groupings for cotton rats and roof rats. Abundance of rodents in Experimental ’06 sites should have been less than that in all other sites if owls were biological contro llers of rodents..........................42 3-5 Results of Dunnett’s tests comparing abundance estimates of cotton rats derived from PVA model scenarios involving various densities of Barn Owls, and incorporating two harvesting intensities and two levels of cotton rat reproduction (A, 75% breeding; B, 45% breeding), to abundances derived from models with no owl predation. Simultaneous 95% confidence intervals that do not encompass zero indicate significance.......................................................................................................... .43 4-1 Model-fit for nest success analysis of 93 Barn Owl nests in the Everglades Agricultural Area, 2004-2007............................................................................................60 4-2 Model-averaged parameter estimates and odds ratio values of predictor variables of Barn Owl nest success in the Ever glades Agricultural Area, 2004-2007..........................60 4-3 Model-fit for analysis of survival of 84 individual Barn Owl chicks in the Everglades Agricultural Area.............................................................................................................. .61 4-4 Model-averaged parameter estimates and odds ratio values for predictor variables of survival of 84 individual Barn Owl chic ks in the Everglades Agricultural Area..............61

PAGE 7

7 LIST OF FIGURES Figure page 2-1 Mean grams of sugarcane consumed per gram of body mass in 24 hrs for cotton rats, roof rats, and rice rats. Solid symbol s = males, hollow symbols = females.....................21 2-2 Linear regression analyses testing if grams of sugarcane consumed per gram of body mass is related to mean air temperature during feeding trials...........................................22 3-1 Everglades Agricultural Area and locations of sites used for experimental manipulation (M1, M2, M3) and control (C1, C2, C3). Dots represent Barn Owl nest boxes installed prior to 2005, pluses represent boxes installed at the end of 2005 for this study…...…...44 3-2 Mean abundance estimates ( SE) for cotton rats, roof rats, and rice rats derived from Huggins closed capture mode ling using Program MARK................................................45 3-3 Peak abundance estimates of cotton rats in a single sugarcane field derived from PVA model scenarios simulating various densities of occupied Barn Owl nest boxes with two levels of predation intensity (1 & 3 rats consumed per owl per month) and two levels of cotton rat reproductive rates (75% and 45% breeding)................................46 4-1 Location of the Evergl ades Agricultural Area...................................................................62 4-2 Daily survival rates ( 95% CI) for Barn Owl nests A) when sugarcane was present and absent within 1500 m, B) during incubation and hatchling nest phases, and C) when roads were present or absent within 1500 m............................................................63 4-3 Daily survival rates ( 95% CI) of individual Barn Owl chicks by age of chick when sugarcane was present and absent......................................................................................64 4-4 Growth curves depicting mean mass, tarsus length, and wing chord length ( SE) for Barn Owl nestlings............................................................................................................. 65

PAGE 8

8 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy ARE BARN OWLS ( Tyto alba ) BIOLOGICAL CONTROLLERS OF RODENTS IN THE EVERGLADES AGRICULTURAL AREA? By Jason Michael Martin May 2009 Chair: Lyn Branch Major: Wildlife Ecology and Conservation Biological control is defined as the use of living organisms to reduce abundance of pest species. The ability of vertebrate predators to control vertebrate pests is limited, however the potential of vertebrates to reduce prey abundance is maximized when predators are capable of reproducing rapidly as prey abundance increases, when their movement is not restricted by territoriality, and if they have high foraging effi ciency. Barn Owls may be effective biological controllers of rodents because they exhibit these traits. Nest boxes have been installed throughout the Everglades Agricultural Area (EAA) in south Florida to increase Barn Owl abundance with the intent of reducing pest rodents in sugarcane fields. This program has successful increased the regional abundance of ow ls, but it is unknown if areas with more owls have lower rodent abundances than areas without owls. Nest boxes are useful tools for expanding nest site availability for many species of birds. However, birds may be drawn to low quality sites by poorly placed boxes. Both immediate quality and long-term stability of habitat must be considered when locating nest boxes or they may act as ecological traps, resulting in low reproductive fitness for the population. When most Barn Owls in the EAA select nesting locations, harvesting activities have not begun and rodents, the primary prey of the owls, are abundant in crop fields. However, harvest occurs adjacent to

PAGE 9

9 most active nest boxes before nesting is completed and therefore most nests are exposed to a rapid decline in food availability. This may have negative consequences to Barn Owls in the form of lowered nest success. This study indicates that Barn Owls are not capable of reducing rodent abundance. Reproductive capacity of rodents far outpaces rem oval of rodents via predation. Furthermore, nest survivorship for Barn Owls was lower following sugarcane harvest when compared preharvest conditions. Long-term data is needed to test for the actual occurrence of an ecological trap, however this study demonstrates that a mechanism for a trap exists. If crop harvest does not result in lowered fitness of Barn Owl populations, then nest boxes may be beneficial conservation tools for Barn Owls.

PAGE 10

10 CHAPTER 1 INTRODUCTION Biological control is defined as the use of living organisms to control pest populations (Sweetman 1958; Caltagirone 1981). In predator-pre y systems, control occurs when prey density is reduced to a level below which it would occur in the absence of predation, and fluctuations in prey abundance are regulated around this suppressed density over time (Potts 1980; Dempster 1983; Hassel 1985). Control of agricultural pests, defined as organisms that inflict unacceptable economic damage (Horn 1988), is achieved when crop loss is mitigated to a tolerable level (Mumford and Norton 1984). However, before socioeconomic management goals can be set, the overall ability of a predator to reduce abundance of the targeted species must be demonstrated. Therefore, in this study I define control as any measurable predator-induced reduction in abundance of prey. The potential of predators to reduce abundance of crop-destroying pests has been recognized for decades (e.g., Shelford 1942; Howard 1967; Rice and Wilde 1988; Borkhataria et al. 2006). A review of studies conducted from 1960-2001 concluded that inundative releases of generalist predators into crop plots resulted in significant pest reduction in most instances; however, in all of these cases, predators and prey were invertebrates (Symondson et al. 2002). Most information pertaining to the ability of vertebrate predators to decrease abundance of vertebrate pests supports the contention that the c ontrolling ability of predators in these situations is limited, although evidence in this regard often lacks experimental rigor (Davis et al. 1976; Newsome 1990). Barn Owls ( Tyto alba ), generalist predators that prey primarily on rodents (Marti 1992, Taylor 1994), may serve as biological controllers of pest rodents in agricultural landscapes. Although this species is declining in parts of North America and Europe because of habitat

PAGE 11

11 alteration and changes in agricultural practices, Barn Owls are common in other parts of their near global distribution (Taylor 1994). Barn Owls are potential biological controllers of rodents because they have a combination of behavioral characteristics and life history traits that many other vertebrate predators lack. Whereas density of carnivorous species often is limited because individuals actively maintain and defend te rritories (Errington 1946, Newton and Marquiss 1991, Lindzey et al. 1994, de Azevedo and Murray 2007), foraging ranges of individual Barn Owls can overlap to a high degree when food and nesting locations are readily available (Ziesemer 1980, Taylor et al. 1988, Taylor 1994). Barn Owls are re latively fecund compared to other vertebrate predators. In portions of the species’ distri bution where climate does not impede reproduction, such as south Florida, Barn Owls are capable of producing two broods each year and fledge 2-6 young per brood (Taylor 1994). Density and location of Barn Owl populations can be manipulated using nest boxes (Duckett 1991, Regisser 1991, Lee 1997); therefore, the collective predation pressure of a population can be mana ged. Adult Barn Owls consume 1-2 rodents per night, and a single pair may eliminate more than 1000 rodents each year (Lenton 1980). When the additional consumption of rodents by nestlings is considered along with the potential to concentrate the collective foraging pressure of a population on an area of interest, the potential of the species to reduce abundance of rodent populations is apparent. Avian species select habitat and nesting locations based on an assessment of habitat quality (Petit and Petit 1996, Muller et al. 1997, Martin 1998, Kolbe and Janzen 2002). Ecological traps occur when environmental cues used to assess quality do not provide an accurate gauge of habitat suitability, primarily due to anthropogenic influences (Misenhelter and Rotenberry 2000, Schlaepfer et al. 2002, Battin 2004, Weldon and Haddad 2005). An integral component of habitat suitability is temporal stability. Organisms must predict long-term

PAGE 12

12 suitability of habitat based on an immediate assessment of quality and do not have the ability to reject good habitat that will be altered in the future. Temporal instability in habitat quality may be especially consequential when selecting breeding sites (Orians and Wittenberger 1991). For example, grassland birds often establish nests in croplands because the structure of these systems mimics that of natural grassland habitat and becau se of the ample food supplies available in these areas. However, nests in croplands often are destroyed by harvesting activities (Best 1986, Bollinger et al. 1990). The decoupling of cue, quality, and temporal stability of habitat can lead to lowered reproductive success of individuals, fa ilure of nests, and extirpation of populations (Kokko and Sutherland 2001, Shochat et al. 2005). Nest boxes are useful tools for expanding nest site availability for many cavity nesting bird species (e.g., Willner et al. 1983, Barry 1992). However, birds may be drawn to low quality sites by poorly placed boxes (e.g., Semel and Sherman 2001). Nest boxes may increase the perceived attractiveness of a site to a species while the actual quality of the site remains unchanged, thus resulting in the nest boxes potentially acting as ecological traps. Furthermore, boxes may be placed in areas that are considered by humans to be high quality based on an immediate assessment of surrounding habitat, but these sites may lack the temporal stability of resources required by the targeted species. Provision of artificial nesting structures may have negative consequences if the wider ecological context of this management action is not fully considered. In the Everglades Agricultural Area (EAA) of south Florida, abundance of Barn Owls historically has been limited because trees large enough to contain hollow spaces suitable for nesting are not common and buildings used by owls for nesting and roosting are disappearing rapidly because of hurricanes and general disuse. Cotton rats ( Sigmodon hispidus ), roof rats

PAGE 13

13 ( Rattus rattus ), and rice rats ( Oryzomys palustris ) are common in the EAA and all are capable of inflicting considerable damage to sugarcane crops (Walsh et al. 1976, Holler et al. 1981, Lefebvre et al. 1989, Martin et al. 2007). Over the past 20 years, approximately 200 Barn Owl nest boxes have been installed throughout the EAA to enhance the regional owl population with the intent of these raptors serving as biological cont rollers of pest rodents. At least two-thirds of the boxes are used by owls during the nesting season (late Aug – June), and all appear to have been used for nesting at least once (Martin, pers obs.). Localized densities in the EAA can reach 18 pairs/5 km2 (Martin, pers. obs.), whereas densities in other parts of this species’ distribution range from 0.7 pairs/5 km2 to 2.6 pairs/5 km2 (Taylor et al. 1988, Altwegg et al. 2003). I hypothesized that increased density of Barn Owls in the Everglades Agricultural Area (EAA) would result in lowered abundances of cotton rats, roof rats, and rice rats. To test this hypothesis, I first estimated abundance of each species over one year in six locations in the EAA that had very low densities of owls. I then randomly selected three of these areas and inflated owl densities in them by installing nest boxes. Following this, I re-assessed rodent abundance over one year and compared preand post-treatment abundances for each species. I also performed a population viability analysis to model the density of barn owls required to produce a measurable decline in abundance of cotton rats. I also hypothesized that survivorship of both Barn Owl nests and individual chicks would be lower following sugarcane harvest compared to when standing sugarcane was present adjacent to nests, primarily because of the extreme crash in rodent abundance that follows harvest. To test this hypothesis, I monitored the fate of nests and individual chicks within a subset of nests, and calculated daily survival rates for preand post-harvest conditions. I also measured and compared the growth rate of the individual chicks over time.

PAGE 14

14 CHAPTER 2 QUANTIFYING DAMAGE POTENTIAL OF THREE RODENT SPECIES ON SUGARCANE Introduction The Everglades Agricultural Area (EAA), located primarily in western Palm Beach County, FL, consists of approximately 160,000 ha (400,000 acres) of sugarcane cropland. Roughly half of the nation’s cane sugar originates in the EAA (Baucum et al. 2006). Three species of rodent, cotton rat ( Sigmodon hispidus ), roof rat ( Rattus rattus ), and rice rate ( Oryzomys palustris ), thrive in this agricultural landscape. Cotton rats are the most abundant of the three species (Doty 1960, Samol 1972b, Holler et al. 1982). Although density of cotton rats may vary among fields because of presence of predators and rodent control programs, this species is a habitat generalist (Whitaker and Hamilton 1998) that occurs throughout the entire region. Roof rats typically reside in fields, as well as near human structures (Whitaker and Hamilton 1998). They are likely to have a distributi on as wide as that of cotton rats in the EAA; however, roof rats are less fecund so their populations recover more slowly following perturbations and their peak densities are likely to be lower. Rice rats primarily are limited to wet, marshy areas (Whitaker and Hamilton 1998), so their distribution in the EAA is restricted. All of these species have been implicated in the destruction of sugarcane crops (Garlough 1938, Doty 1960, Samol 1972a, Samol 1972b, Walsh et al. 1976, Lefebvre et al. 1978, Holler et al. 1982), but an empirical investigation of the relative amount of damage caused by each is lacking (Samol 1972b). Current pest control effo rts focus on reducing the number of rodents in general across the entirety of the region by applying rodenticides broadly (Abarca 1981, Lefebvre et al. 1985, Montague et al. 1990). However, due to variation in habitat specificity and reproductive capacity described above, both density of individual species and overall species composition of rodents varies among fields (Doty 1960, Holler et al. 1982, Lefebvre et al. 1982,

PAGE 15

15 Lefebvre et al. 1989). If certain species are responsible for a disproportionately large amount of damage to sugarcane crops, the efficiency of management efforts may be improved by focusing pest control efforts on areas where the most pr oblematic species are most abundant. This study investigates and compares the amount of sugarcane that cotton rats, roof rats, and rice rats are capable of consuming in 24 hrs using a feeding trial experiment. Methods Feeding trials were conducted from 12 Mar 4 May 2005, 12 Oct 2005 4 Jan 2006, and 15 Dec 2006 9 Feb 2007. Individuals of each speci es were opportunistically captured using single-door, collapsible Tomahawk live traps (40.5 x 12.5 x 12.5 cm, Tomahawk Live Trap Co., Tomahawk, WI) in a single sugarcane field of cultivar ‘CP 80-1743’ (Deren et al., 1991) at the Everglades Research and Education Center (ERE C), Belle Glade, FL. Traps were opened and baited with rolled oats between 1700-1800 hrs, and then checked the following morning between 0700-0800 hrs. Rats were removed from occupied traps using a cloth bag, and species, sex, and mass were recorded. Individuals then were placed in separate 46 x 46 x 46 cm wire mesh cages with a pre-weighed, two-internode long section (approximately 30 cm in length) of ‘CP 80-1743’ sugarcane stalk (mean mass = 160.8 g, range: 103.1 – 280.3, SE = 2.0). No more than 6 rats were captured and tested simultaneously due to limited cage availability. All individuals captured were used in the experiment; therefore, the sample population represented a cross section of the sex and age classes in the field. Approximately two heaping teaspoon of rolled oats were added to each cage to simulate naturally occurring alternative food sources. An additional pre-weighed stalk section was placed in an empty cage for each group of rats tested to serve as a control for mass lost from stalks due to desiccation. All stalk sections were cut from the bottom six internodes of mature stalks harvested from the same field where rats were captured. Cages were placed in the breezeway of a barn away from human activity and were

PAGE 16

16 separated from each other with sheets of paper. After 24 hrs, rats were released and all stalks were weighed again. Mass lost from the contro l stalk was subtracted from mass lost from experimental stalks to determine total grams of sugarcane consumed by each rat (g sugarcane/g rat/24 hrs). Mean air temperature was obtained for each trial period from the Florida Automated Weather Station located at EREC. Before data analyses were conducted, data were tested for normality, a square root transformation was applied when data violat ed the assumption of normal distribution, and outlying data points were removed (Proc Univariate, SAS Institute, 2003). A fixed effect ANCOVA was used to examine if species, sex, mass, mean air temperature, and interactions of these variables influence the rate of sugarcane consumption by rats (Proc GLM, SAS Institute, 2003). Mean grams of sugarcane consumed per gram of body mass were calculated for each sex of each species. Finally, linear regression analyses were used to explore the relationship between mean air temperature and grams of sugarcane consumed per gram of body mass for each species (Proc REG, SAS Institute, 2003). A significance level of = 0.05 was used for all statistical tests. Data are presented as means SE. Results During this study, 114 cotton rats, 60 roof rats, and 42 rice rats were captured and tested (Table 2-1). ANCOVA indicated that species, sex, and an interaction between species and mean air temperature influenced the amount of sugarcane consumed by rats during the trials. Sex and the species*mean air temperature interaction term fell just short of statistical significance, but I still considered them relevant factors (Table 2-2) Rate of sugarcane consumption (g sugarcane/g rat/24 hrs) was not related to body mass of individuals for all species, and not related to sex of cotton rats (males: 0.39 0.03 g; females: 0.34 0.03 g) or roof rats (males: 0.25 0.04 g;

PAGE 17

17 females: 0.26 0.04 g). Male rice rats, however, consumed a significantly higher percentage of their body mass than female rice rats (males: 0.49 0.08 g; females: 0.20 0.06 g). Cotton rats consumed a higher percentage of their body mass than roof rats. Grams of sugarcane consumed per gram of body mass by female rice rats did not differ from roof rats, but male rice rats consumed a significantly higher percentage of their body mass than any other group (Fig. 2-1). Regression analysis showed a negative relationship between grams of sugarcane consumed per gram of body mass and mean air temperature for rice rats, but there was no relationship between these variables for cotton rats or roof rats (Fig. 2-2). Discussion Consumption rate of sugarcane was similar across all body sizes within each of the three species of rodents. Thus, amount of sugarcane consumed in an area can be estimated by multiplying the following three parameters: consumption per gram of body mass of the species, mean mass of individuals, and density. For example, based on the mean mass of cotton rats that were captured in 34 locations throughout the EAA in 17,000 trap nights in 2005-2006 (Martin, unpublished data) and population density estimates from a previous study in the EAA (Lefebvre et al., 1982), our experimental data predict that cotton rats may consume 41.1 – 198.6 kg of sugarcane stalk per ha, or 7,056 – 32,442 metric tons across the entire EAA, during the four months prior to harvest when the annual rodent density cycle is at its peak. Individual cotton rats are capable of consuming at least as much sugarcane as roof rats and rice rats. Because it is the most common of the three rodent pest species, my experimental data suggest that cotton rats collectively are responsible for the greatest proportion of rodent damage to sugarcane crops in the EAA. Roof rats consume as much sugarcane as individual cotton rats; however, because they are less abundant, their collective contribution to sugarcane destruction in the EAA should be less. Rice rats probably consume less sugarcane across the

PAGE 18

18 EAA than cotton rats and roof rats because their distribution is restricted to fields with high soil moisture. However, rice rats may collectively cause more damage to sugarcane in wetter fields where they are more common. The large surface to volume ratios of small animals often result in high metabolic rates and a need for greater amounts of food in colder temperatures in order to maintain body temperature (Geiser et al. 2006). Rice rats, the smallest of the three rodents, exhibited thermoregulatory feeding behavior during the feed ing trials. Fields occupied by rice rats may experience spikes in damage during periods of colder temperatures. Actual within-field patterns of rodent damage may vary from the results of this experiment because of conditions not measured in this study, such as predator density, moon phase, precipitation, or the availability of alternative food sources. Also, the damage that rodents inflict on sugarcane extends beyond the amount of cane consumed. Even small breaks in the rind of stalks resulting from the gnawing of rode nts may result in reduced nutrient flow through the plant and secondary fungal infections which further reduce sucrose yield (Abarca 1981). The physical structure of sugarcane fields may infl uence occurrence of rodent damage. Sugarcane stalks sometimes bend or break when they are tall. This lodging results in stalks leaning against each other, growing at an angle instead of ver tically, or lying flat on the ground. Rodents are excellent climbers and are not limited to damaging only the portions of vertical stalks that they can reach from the ground; however, lodged stalks provide easier access to the entire length of the plant. Fields with greater amounts of lodgi ng may have greater amounts of rodent damage. Practically speaking, the overall amount of sugarcane crop damage is unlikely to vary with composition rodent communities within fields. Cotton rats, roof rats, and rice rats all consume sugarcane, and the cumulative damage inflicted by rodents will be great when rodent

PAGE 19

19 densities are high. The amount of damage to crops will increase as the density of rodents increases, regardless of which particular species are present. Density of rodents in sugarcane fields should be monitored periodically to identify areas where rodent control is most critical. Elimination of rodent refugia habitat, such as overgrown fallow fields and weedy drainage ditch banks, may help limit regional rodent abundance. These areas provide safe havens for rodents after agricultural fields have been harvested. Rodent abundance may be further reduced by making farms more attractive to predators. An integrated pest management approach, which incorporates ecologically friendly management techniques such as these, may lessen need for chemical rodenticides while maintaining or enhancing overall control of rodents.

PAGE 20

20 Table 2-1. Mass of rodents tested in feeding trials. Species n Mean mass SE (g) Range Cotton rat (male) 72 120.5 4.536-208 Cotton rat (female) 42 111.4 5.350-202 Roof rat (male) 36 182.3 6.980-270 Roof rat (female) 24 157.1 5.590-208 Rice rat (male) 33 81.8 3.852-132 Rice rat (female) 9 58.1 1.950-66 Table 2-2. ANCOVA examining the influence of species, sex, mass, and mean air temperature on grams of sugarcane consumed per gram of body mass by rats during feeding trials. Sugarcane consumed (g) / body mass (g) F p s p ecies 3.250.04 sex 3.520.06 species*sex 0.850.43 mass 0.900.35 species*mass 1.820.17 mean temp 0.700.84 mean temp*species 2.600.08 mean temp*mass 0.330.57 mean temp*species*mass 1.150.32

PAGE 21

21 0 0.1 0.2 0.3 0.4 0.5 0.6 00.511.522.533.544.5Sugarcane consumed (g) / body mass (g) (mean SE) Figure 2-1. Mean grams of sugarcane consumed per gram of body mass in 24 hrs for cotton rats, roof rats, and rice rats. Solid symbols = males, hollow symbols = females. Cotton rat Roof ratRice rat

PAGE 22

22 Figure 2-2. Linear regression analyses testing if grams of sugarcane consumed per gram of body mass is related to mean air temperature during feeding trials. 0.0 0.2 0.4 0.6 0.8 1.0 1.2 51015202530R2=0.02 df=1, 113 P=0.13 0.0 0.2 0.4 0.6 0.8 1.0 51015202530Sugarcane consumed (g) / body mass (g)Roof rat R2=0.004 df=1, 60 P=0.162 0.0 0.4 0.8 1.2 1.6 2.0 51015202530Rice rat Air Temperature (C) R2=0.20 df=1, 41 P=0.003 Cotton rat

PAGE 23

23 CHAPTER 3 AN ASSESSMENT OF BARN OWLS AS BIOLOGICAL CONTROLLERS OF PEST RODENTS Introduction Biological control is defined as the use of living organisms to control pest populations (Sweetman 1958; Caltagirone 1981). In predator-pre y systems, control occurs when prey density is reduced to a level below which it would occur in the absence of predation, and fluctuations in prey abundance are regulated around this suppressed density over time (Potts 1980; Dempster 1983; Hassel 1985). Control of agricultural pests, defined as organisms that inflict unacceptable economic damage (Horn 1988), is achieved when crop loss is mitigated to a tolerable level (Mumford and Norton 1984). However, before socioeconomic management goals can be set, the overall ability of a predator to reduce abundance of the targeted species must be demonstrated. Therefore, in this study I define control as any measurable predator-induced reduction in abundance of prey. The potential of predators to reduce abundance of crop-destroying pests has been recognized for decades (e.g., Shelford 1942; Howard 1967; Rice and Wilde 1988; Borkhataria et al. 2006). A review of studies conducted from 1960-2001 concluded that inundative releases of generalist predators into crop plots resulted in significant pest reduction in most instances; however, in all of these cases, predators and prey were invertebrates (Symondson et al. 2002). Most information pertaining to the ability of vertebrate predators to decrease abundance of vertebrate pests supports the contention that the c ontrolling ability of predators in these situations is limited, although evidence in this regard often lacks experimental rigor (Davis et al. 1976; Newsome 1990). The ability of vertebrate predators to decrease abundance of vertebrate pests may be limited for several reasons: 1) removal rate of prey by predators usually is far outpaced by

PAGE 24

24 reproductive capacity of prey species (Pearson 1966), 2) predation often is compensatory rather than additive (Errington 1967; Bailey 1984), 3) predators may respond numerically to increases in intraspecific competition or declines in prey abundance by reducing reproductive output (Solomon 1949; Bayliss and Choquenot 2002), and 4) predators may respond functionally to changes in prey abundance by lowering per capita rate of food consumption or by shifting to alternative prey (Solomon 1949; Quinn and Cresswell 2004). The potential of vertebrate predators to reduce and regulate prey abundance is maximized when predators are capable of reproducing rapidly as prey abundance increases, when their movement is not restricted by territoriality, if they have high foraging efficiency, and if there is an alternative food source available to sustain their population when primar y prey become scarce (MacArthur 1955; Davis et al. 1976; Hagen et al. 1976; Korpimaki and Norrdahl 1989; Symondson et al. 2002). If these conditions are met, prey may be driven to near extirpation (Sinclair et al. 1990). Intraspecific competition for limited food resources often restricts predator density (Errington 1946; Votier et al. 2007; Bonal and Apar icio 2008); however, if predator density can be increased by altering a limiting resource not related to food (e.g., potential nesting sites), the collective rate of prey removal within a given area may increase. This management strategy has produced mixed results (e.g., Kale 1968; Meserve et al. 1993; Kay et al. 1994; Norrdahl and Korpimaki 1995). The success of habitat manipulation programs in enhancing predation is dependent on the ecology of both the predator and the targeted prey species, as well as the complexity of the surrounding ecosystem. Barn Owls ( Tyto alba ), generalist predators that prey primarily on rodents (Marti 1992; Taylor 1994), may serve as biological controllers of pest rodents in agricultural landscapes. Although this species is declining in parts of North America and Europe because of habitat

PAGE 25

25 alteration and changes in agricultural practices, Barn Owls are common in other parts of their near global distribution (Taylor 1994). Barn Owls are potential biological control agents because they have a combination of behavioral characteristics and life history traits that many other vertebrate predators lack. Whereas density of carnivorous species often is limited because individuals actively maintain and defend terr itories (Errington 1946; Newton and Marquiss 1991; Lindzey et al. 1994; de Azevedo and Murray 2007) foraging ranges of individual Barn Owls can overlap to a high degree when food and nesting locations are readily available (Ziesemer 1980; Taylor et al. 1988; Taylor 1994). Barn Owls are relatively fecund compared to other vertebrate predators. In portions of the species’ distri bution where climate does not impede reproduction, such as south Florida, Barn Owls can produce two broods each year and fledge 2-6 young per brood (Taylor 1994). Density and location of Ba rn Owl populations can be manipulated using nest boxes (Duckett 1991; Regisser 1991; Lee 1997); therefore, the collective predation pressure of a population can be managed. Adult Barn Owl s consume 1-2 rodents per night, and a single pair may eliminate more than 1000 rodents each year (Lenton 1980). When the additional consumption of rodents by nestlings is considered along with the potential to concentrate the collective foraging pressure of a population on an ar ea of interest, the potential of the species to reduce abundance of rodent populations is apparent. Several previous studies have attempted to artificially increase Barn Owl density with the intent of reducing abundance of pest rodent s (e.g., Muoz and Mura 1990; Lee 1997; Moore et al. 1998; Potter 2004). In these studies, nest boxes proved to be an effective means of manipulating Barn Owl density and targeted pest species comprised a significant portion of the diet of the owls. However, none of these studies quantified pest populations to determine if owl

PAGE 26

26 predation resulted in reduced pest abundance and therefore provided no evidence regarding the use of Barn Owls as biological controllers (Marsh 1998). Research Design I hypothesized that increasing density of Barn Owls in the Everglades Agricultural Area (EAA) would result in lowered abundances of cotton rats, roof rats, and rice rats. To test this hypothesis, I used a completely randomized study design. Six areas in the EAA were selected that initially had very low densities of owls. Rodent abundances were estimated in each area over one year. I then randomly selected three of these areas and inflated owl densities in them by installing nest boxes. Following this, I re-assessed rodent abundance over one year and compared preand post-treatment abundances for each species. Finally, I performed a population viability analysis to model the density of barn owls required to produce a measurable decline in abundance of cotton rats in an individual sugarcane field and across the entire EAA. I predicted that abundances of all species would be lower following inflation of owl density when compared to pre-manipulation abundances or abundance estimates from control areas that had no owls. I assumed that cotton rats, roof rats, and rice rats comprised a large proportion of the diet of Barn Owls in the EAA. To test this assumption, I quantified the regional owl diet by examining the content of owl pellets. I also assumed that density of other predators of rodents, such as diurnal raptors, were constant throughout all areas. Alternatively, rodent abundance may differ among sugarcane fields in the EAA because of variation in sugarcane variety (Barry Glaz, USDA Sugarcane Fi eld Station, Canal Point, FL, pers. com.) or rotational age of fields (Samol 1972), however I controlled for both of these factors when selecting locations to sample rodents. I speculated that if rodent abundances were not significantly lower following inflation of owl density, this would be because reproductive capacity of all three species exceeded the rate of removal of individual rodents via predation.

PAGE 27

27 Methods Study Location and Species I conducted this study in the Everglades Agricultural Area (EAA), a farming region in south Florida where sugarcane is the primary cr op (Fig. 3-1). The EAA is divided by a network of drainage ditches into fields measuring approximately 200 x 800 m. These ditches may provide several critical resources for rodents, including refugia habitat following harvest of sugarcane and potential movement corridors (Robertson et al. 1995). Several cultivars of sugarcane are grown simultaneously across the EAA in 3-6 year rotational cycles resulting in a landscape mosaic of cultivars and rotational ages. Prior to the initial rotation year, referred to as “plant cane”, soil is disked in December and January before sugarcane is planted. This process may destroy rodent burrow systems, thereby displacing localized populations. Fields are harvested following a 10-14 month growing season and remaining stalk stubble is left in the ground to re-sprout during subsequent rotational year s, referred to as “ratoons”. Soil disturbance during ratoon years is minimal. Cotton rats ( Sigmodon hispidus ), roof rats ( Rattus rattus ), and rice rats ( Oryzomys palustris ) are common in the EAA and all are capable of inflicting considerable damage to sugarcane crops (Walsh et al. 1976; Holler et al. 1981; Lefebvre et al. 1989; Martin et al. 2007). Damage to sugarcane plants may occur more frequently in older ratoon fields than in fields that are earlier in the rotation cycle (Samol 1972). Barn Owls occur in grasslands, agricultural fields, and other open habitat throughout most of the world, and feed primarily on rodents (Marti 1992; Taylor 1994). The use of suitable foraging habitat by this species is limited by availability of nesting cavities (Marti 1992). Because rodent abundance in the EAA is high, the availability of potential nest sites likely is the primary limiting factor of Barn Owl density in the region. Trees large enough to contain hollow spaces suitable for nesting are not common. Furthermore, buildings used by owls for nesting and

PAGE 28

28 roosting are disappearing rapidly throughout the region because of hurricanes and abandonment by farmers. Over the past 20 years, approximate ly 200 Barn Owl nest boxes have been installed in the EAA to enhance the regional owl population with the intent of reducing pest rodent abundance. Throughout this study, owls nested and roosted in at least two-thirds of these boxes (Martin, pers. obs.). Anecdotal reports describe rodent declines and less crop damage in areas of the EAA that are occupied by Barn Owls; however, prior to this study, quantitative data had not been collected to confirm these observations. Study Design Study sites and experimental manipulation of owl density: I quantified abundance of cotton rats, roof rats, and rice rats in six areas in the EAA and subsequently manipulated Barn Owl density in three of these areas chosen at random. All areas had extremely low owl densities at the beginning of my study in 2005 because of a lack of potential nesting and roosting locations. In late autumn 2005, following one year of surveys of rodent abundance, I installed 25 nest boxes in each of the three experimental areas to inflate Barn Owl abundance for the following year (M1, M2, M3; Fig. 3-1). I observed 10-16 nests ( x = 14.5, SD = 2.3) in the boxes at each site during the 2006 nesting season. In 2006, I conducted a second round of rodent surveys in the experimental owl manipulation areas and in the three areas with no nest boxes that served as controls (C1, C2, C3; Fig. 3-1). All study areas were 10-km2, representing a typical home range size of Barn Owls (Hegdal and Bl askiewicz 1984; Marti 1992; Taylor 1994), and were spaced at least 5 km from other study areas to ensure independence of predation pressure by owls on local rodent communities. Rodenticides were not applied to any areas involved in this study. Protocol for estimating rodent abundance: I used capture-mark-recapture techniques to estimate abundance of cotton rats, roof rats, and rice rats in all study areas during the 2005 and

PAGE 29

29 2006 sugarcane growing seasons. Throughout the majo rity of the growing year, sugarcane stalks lodge together resulting in a thick mass of vegetation that likely restricts the majority of predatory interactions between Barn Owls and rode nts to field edges and drainage ditch corridors that separate individual fields. Therefore, all rodent surveys were conducted along drainage ditches. In January 2005, immediately following the 2004 sugarcane harvest, two ditches were selected in each study area. These ditches were located at least 800 m apart, which is more than two times previously described home range sizes and movement distances for cotton rats (Slade and Swihart 1983; Camerson and Spencer 1985; Di fendorfer and Slade 2002; Sulok et al. 2004), roof rats (Whitaker and Hamilton 1988; Pryde et al. 2003), and rice rats (Whitaker and Hamilton 1988). One ditch was located between two plant cane fields and the other between older fields (at least second ratoon). To control for potential influence of sugarcane variety on rodent abundance, surveys were limited to ditches between fields containing commonly grown cultivars (CP80-1743 or CP89-2143). Beginning at a randomly selected end of each ditch, I established two 250-m long transects, one on each side of the ditch, following the second plant row in from the edge of each field adjacent to the ditch. The two transects were separated by approximately 20 m. Two Tomahawk live-traps (40.5 x 12.5 x 12.5 cm, Tomahawk Live Trap Co., Tomahawk, WI) were placed every 10 m along each transect (100 traps per trapping array). For five consecutive days, I baited traps with rolled oats and opened them between 1600 and 1800 h. I checked and closed them the following morning between 0700 and 1000 h. I uniquely marked all cotton rats, roof rats, and rice rats captured by clipping fur, and recorded species, sex, and mass. Immediately

PAGE 30

30 following processing, I released individuals at the capture location. Fur clips were not retained beyond each 5-day trapping session. I conducted three trapping sessions during each of two growing seasons (2005, 2006), with the six study areas surveyed consecutively in a random order. During each trapping session, I trapped the two ditches within each study area concurrently. The initial set of surveys took place from 24 Jan – 26 Mar 2005 in recently harvested fields. Vegetative cover on ditch banks was lacking during this early field gr owth phase due to mowing and/or herbicide application. I then conducted a second series of mid-growth phase surveys (23 May – 23 July 2005) in the same order as the early growth phase surveys. At this time, fields contained 1 3 m tall rows of cane, and grasses, forbs, and sedges had begun to grow along ditch banks. Finally, I conducted a series of late growth phase surveys prior to harvest from 26 Sept – 10 Dec 2005 when sugarcane was 3.5 4.5 m tall and ditch banks were thickly vegetated. Trapping was not conducted from 23 Oct – 22 Nov 2005 because of Hurricane Wilma. In 2006, I selected two new field ditches in each study area and trapped them in a random order during the same three phases of sugarcane growth (early: 6 Feb – 8 Apr, mid: 29 May – 29 July, late: 25 Sept – 25 Nov). Barn Owl dietary analysis: Barn Owls typically swallow prey whole and later regurgitate compacted pellets of indigestible bones and fur (Taylor 1994). Examination of prey remains found in these pellets is a reliable indicator of diet (Errington 1932; Raczynski and Ruprecht 1974; Dobson and Wexlar 1979). To determine if cotton rats, roof rats, and rice rats comprised a significant portion of the Barn Owl diet in the EAA, I examined the contents of 1676 pellets collected between May 2004 and Feb 2007 from 15 buildings located throughout the area known to be used by owls for nesting or roosting. Because remains of larger prey items may be split among several pellets, only skulls were used to quantify prey base to avoid double

PAGE 31

31 counting individuals. All small mammal skulls found were identified to species using reference specimens provided by the Florida Museum of Natural History. Rabbits were identified to genus (Sylvilagus), frogs and toads were identified to order (Anura), and birds were identified to class (Aves). The total number of skulls belonging to each species or taxonomic group was counted, and the relative percentage of each was calculated to determine the proportion of each in the diet of the EAA Barn Owl population. Population viability analysis of cotton rats: I performed a population viability analysis (PVA) using Vortex 9.94 (Lacy et al. 2008) to model the density of Barn Owls required to produce a measurable reduction in abundance of cotton rats, the most abundant rodent in the EAA (Doty 1960, Samol 1972b, Holler et al. 1982), at the scale of a single sugarcane field. I also modeled the relative impact of two levels of predation intensity and two levels of cotton rat reproductive rates in this hypothetical system. I initially modeled a population of cotton rats over one year using minimum population size and carrying capacity estimates derived from data collected during this study from areas that did not have nest boxes, as well as life history parameters previously documented for the species (Montague et al. 1990; Whitaker and Hamilton 1998). I then adjusted these parameters so that the population growth curve predicted by the model matched data that I collected in areas without nest boxes (Table 3-1). I assumed that rates of dispersal into and out of the population were balanced. I created scenarios simulating the cotton rat population within the foraging territories of owls surrounding various numbers of owl boxes: 1 box, 4 boxes, 6 boxes, 8 boxes, and 10 boxes. I assumed that all boxes in each scenario were occupied, and each had 2 adults and 4 chicks in them for 5 months (assuming double broods) and only 2 owls for the remainder of the year. Therefore, the mean number of owls per month represented in each scenario was 4, 16, 24,

PAGE 32

32 32, and 40 respectively. In the EAA, there are approximately 45 fields within 1500 m (the size of an owl home range) of each nest box. Individual Barn Owls consume an average of 1.5 rodents each night (Taylor 1994). If predation pressure was distributed equally throughout home ranges, each owl would consume one rodent per field per month. Therefore, I first modeled the impact of predation on abundance of cotton rats for each owl density scenario by setting the harvest rate to one cotton rat per owl per month. The use of certain areas by raptors may be enhanced by providing artificial perches (Hall et al. 1981; Reinert 1984), therefore it may be possible to increase the removal rate of prey by individual owls in targeted areas. I simulated an increased concentration of predation on the cotton rat population by inflating the harvest rate to three rats per owl per month for each nest box density. In addition to increased adult mortality, abundance of wildlife populations may be impacted by changes in reproduction (Krebs and Meyers 1974). To examine how a reduced reproductive rate in cotton rats would influence the impact of Barn Owl predation on cotton rat abundance, I repeated the simulations described above using a reduced percentage of reproductive females (from 75% to 45%). A total of 100 simulations were completed for each scenario. Mean abundances ( SD) were estimated for each scenario at the end of one year. Data analysis I estimated abundance of cotton rats, roof rats, and rice rats for each field ditch using Huggins closed capture modeling in Program MARK (White and Burnham 1999). The conditions of population closure were assumed to be approximated during each 5-night trapping session. Abundance estimates were used as relative indices to compare rodent abundance among areas and over time. To estimate abundance, I first modeled species-specific capture and recapture probabilities. Because habitat structure in sugarcane fields differed substantially during each of

PAGE 33

33 the three field growth phases in which I conducted surveys, I modeled detection probabilities for each rodent species within each field phase separately. Sample sizes often were too small to allow for adequate estimation of ditch-specific detection probabilities by species and phase, so capture-mark-recapture data for cotton rats and roof rats each were pooled across all trapping locations for each phase using ditch as a grouping term to estimate pooled detection probabilities. Because rice rats were less common than cotton rats and roof rats, data on rice rats were pooled further by combining early and mid-season field growth phase data. I incorporated heterogeneity in detection probability of individual rodents using sex and age as covariates. Age classifications were determined for each species using mass categorizations from previous studies (cotton rat: <61 g = juvenile, 61-110 g = subadult, >110 = adult (Sauer and Slade 1985); roof rat: <71 g = juvenile, 71-120 g = subadult, >120 g = adult (Tamarin and Malecha 1972); rice rat: <30 g = juvenile, 30-50 g = subadult, > 50g = adult (Forys and Dueser 1993)). I constructed models for each species within each field phase that included no covariates, as well as age, sex, age and sex, and an interaction of age and sex on capture probability (p) and recapture probability (c) for scenarios where detection probability was constant (p = c) and where there was a behavioral response to trapping (p c). Behavioral response models were not estimable for cotton rat data from the early field growth phase, so in this instance I constructed only constant detection probability models. I used an AICc approach in Program MARK to rank competing models. I then used pooled detection probabilities to derive abundance estimates for each species occupying each ditch for each model. Although using pooled detection probabilities derived from grouped data may bias abundance estimates for individual trapping arrays, this approach produces more accurate estimates than simply using number of individuals captured (Conn et al. 2006; Cooch

PAGE 34

34 and White 2008). I calculated final abundance estimates (N ˆ) for each species occupying each field ditch during each field growth phase by m odel averaging abundance estimates derived from all models of data collected from the respective ditches (Cooch & White, 2008). The influence of covariates on N ˆ can not be modeled directly in Program MARK (White 2002). Therefore, to examine the influence of Barn Owl density on abundance of cotton rats, roof rats, and rice rats, I used a repeated measures ANOVA with field ditch as the subject variable and field phase as the repeated variable (Proc Mixed, SAS 2003). Appropriate covariance structures were selected using best f it statistics. To account for varying precision of N ˆ values, I weighted each estimate by the inverse of its variance using the ‘weight’ statement in SAS. To avoid excluding values of N ˆ= 0 (i.e., ditches where no rodents were captured) from analyses, I set their weights to 0.0001. A significance level of p = 0.05 was used throughout. I initially examined the influence of field age (plant cane vs. older ratoon) on abundance of each rodent species independently of other c ovariates in all study areas. Field age was not significant in any case and therefore was excluded in all further analyses. I then analyzed abundance of each species using the following covariates: field growth phase (early, mid, late), year (2005, 2006), treatment group (experiment al, control), treatment*phase, year*phase, year*treatment, and year*treatment*phase. Leas t square mean estimates of rodent abundance were calculated for all factors. I then performed pair-wise least square mean comparisons using t tests between the following groups: Control ’05 vs. Control ’06, Experimental ’05 vs. Experimental ’06, Control ’05 vs. Experimental ’05, Control ’06 vs. Experimental ’06. To analyze PVA results, I first used an analysis of variance (SAS 2003) to compare final abundance estimates derived from the low and high reproductive rate scenarios with no owl boxes. Using Dunnett’s test, I then performed pa irwise comparisons of abundances derived from

PAGE 35

35 each nest box density scenario and the control estimate derived in the absence of nest boxes for each predation pressure/reproductive rate combination (1 rat/owl/month & 75% rat reproduction, 3 rats/owl/month & 75% rat reproduction, 1 rat/owl/month & 45% rat reproduction, 3 rats/owl/month & 45% rat reproduction). In this test, if the simultaneous 95% confidence interval of a comparison does not encompass zero, then the treatment value is significantly different than the reference value. Results Modeling rodent detection probabilities: All competing models of detection probabilities (i.e., those with AICc < 2) for all species contained age and/or sex covariates for capture and/or recapture likelihoods (Table 3-2). With the exception of early season data for cotton rats, where behavioral response models were omitted, models that incorporated behavioral response to trapping were ranked exclusively as competing models for all species in all seasons. Rodent abundance: Field growth phase was highly significant in determining abundance of cotton rats and roof rats (Table 3-3). Abundance of both species increased as sugarcane in surrounding fields grew, with cotton rats ultimately becoming much more abundant by the late growth phase than roof rats (Fig. 3-2) The three-way interaction of year, field phase, and treatment was significant for cotton rats and roof rats indicating that patterns in abundance varied across years, field phases, and treatments (Table 3-3). However, none of the pair-wise comparisons of rodent abundance among study site groupings pertinent to this study were significant (Table 3-4). Barn Owl dietary analysis: I found 2146 individual prey items in owl pellets. Sixty percent of these were rodents identified as sugarcane pests: cotton rats (42%), roof rats (10%), and rice rats (8%). House mice (Mus musculus) also comprised a relatively large portion of the Barn Owl diet (29%). The remainder consisted of rabbits (5%), birds (4%), round-tailed

PAGE 36

36 muskrats (Neofiber alleni; 1%), southeastern shrews (Sorex longirostris; 0.7%), and toads/frogs (0.3%). Insect remains were found in many pellets, especially in those collected in June and July, but were not quantified. Population viability analysis of cotton rats: The abundance estimate for cotton rats derived from the model with no owls and 45% of female cotton rats breeding was significantly lower than when owls were absent and 75% of female rats were breeding (F = 154.6, df = 1, 198, p < 0.01; Fig. 3-3). When 75% of female rats we re breeding and individual owls harvested 1 rat per month, only scenarios with 8 and 10 boxes (32 and 40 owls respectively) resulted in cotton rat abundances that were significantly less than th e control scenario with no owl predation (Table 3-5, Fig. 3-3). However, when the harvest rate was increased to 3 rats per owl per month with the same percentage of reproductive females, the number of boxes required to produce a measurable decline in cotton rat abundance dropped to 6 (24 owls). When only 45% of female rats were breeding, significant declines in cotton rat abundance were apparent with only 4 nest boxes (16 owls) at both predation intensities. Discussion Barn Owls in the EAA prey heavily on cotton rats, roof rats, and rice rats. However, the results of this study indicate that increasing Barn Owl density alone is unlikely to result in reductions of pest abundance in large agricultural landscapes. Reproductive capacity of rodents likely far exceeds rate of removal of individuals from the population via predation. Cotton rats are capable of producing litters of 10-15 young every month throughout the year, and roof rats and rice rats are only slightly less fecund (Whitaker and Hamilton 1998). Although late season abundance estimates for roof rats and rice rats appeared to decline slightly in the experimental sites following inflation of owl density, these di fferences were not significant. Even if a marginal reduction in rodent abundance actually occurred, this likely would not have resulted in

PAGE 37

37 a measurable reduction in overall crop damage given the potential amount of sugarcane that individual rodents are capable of consuming (Martin et al. 2007). Rodent populations occur at the scale of landscapes, but owl predation is a relatively localized phenomenon restricted primarily to areas near active nest boxes. Depredated rodents within foraging ranges of owls likely are replaced continually by dispersing individuals from the surrounding landscape. Increasing density of Barn Owls across the EAA at the scale required to reduce regional rodent abundance likely is not feasible, as indicated by the population viability analysis of cotton rats. When predation pressure of individual owls is spread equally throughout their foraging territories and the reproductive rate of cotton rats is high, each field in the EAA would need to be exposed to at least 8 occupied nest boxes. This translates to boxes spaced in a 950x950-m grid across the entire EAA (2113 boxes in total) with a 100% occupancy rate. This spacing is only slightly more dense than that of nest boxes in the experimental manipulation sites of this study, and in these sites maximum occupancy of nest boxes was only 64%, although this occupancy rate was based only on one year. Field growth phase significantly influenced abundance of all three rodent species. Rodent populations increased in abundance throughout the sugarcane growing season as food and cover increased and then crashed following harvest. Because cotton rats, roof rats, and rice rats comprise such a large part of the Barn Owl diet in the EAA, declines in these species may result in declines in owl nesting success. Dietary diversity may buffer against declines of a portion of the prey base, however, house mice, the most common non-pest prey item, are much smaller than any of the pest species and do not provide comparable per capita nutritional intake. Additionally, all rabbit remains found in pellets belonged to juveniles that were no larger than

PAGE 38

38 rats, and rabbits, along with the remaining prey items, comprised only a small fraction of the owls’ total diet. Further studies are needed to determine if Barn Owl predation can reduce rodent abundance if combined with other pest control techniques, such as strategically timed application of rodenticides and elimination of rodent refugia habitat following harvest, in an integrated pest management (IPM) approach. IPM seeks to implement a variety of compatible pest control tactics (chemical, biological, and cultural) to reduce abundance of crop-destroying pests (Hoskins et al. 1939, Smith 1978, Kogan 1998). PVA modeling indicated that increasing foraging intensity of individual owls and decreasing the reproductive capacity of rodent populations may be effective components of pest rodent management. The potential of owls to reduce abundance of rodents may increase if preda tion is concentrated on targeted areas, possibly by installing artificial perches in areas where pest control is desired (Hall et al. 1981; Reinert 1984). It is unlikely that this would result in reduced abundance of rodents in the EAA given the widespread nature of the rodent population. However, this approach may be more practical in small-scale agriculture. PVA modeling also indicated that lowering rodent reproductive rates may produce dramatic decreases in abundance, howev er the effectiveness and safety of wildlife contraception as a management tool is not certain (Guynn 1993, Warren 1995, Nettles 1997, Fagerstone et al. 2006). Although inflation of owl density through insta llation of nest boxes does not appear to be an effective means of reducing abundance of pest rodents in large agricultural systems, this management tool may be important for conservation of owls as traditional nest sites of this species become scarce. As in previous studies (e.g., Muoz and Mura 1990; Lee 1997; Moore et al. 1998; Potter 2004), I found that Barn Owl density can easily be manipulated by installing

PAGE 39

39 nest boxes. However, the success of this management technique exhibited in this study may have been greater than that in other areas because of extremely high regional abundance of owls in the EAA that already existed prior to this study. Nevertheless, although Barn Owl management programs involving nest boxes should not expect colonization to the degree that I achieved, nest boxes are a proven management option where increasing abundance of the species is desirable. This study demonstrates the inability of vertebrate predators to serve as biological controllers of vertebrate pests, even when individuals of the species have traits that maximize their ability to reduce abundance of these pests (i.e., relatively fecund, lack of territoriality, high foraging efficiency). Pests are abundant in anthropogenically altered landscapes or become concentrated in specific areas because typical bottom-up controlling mechanisms, such as limited availability of food, do not exist or exhibit only a weak influence on abundance. In agricultural areas, crops may provide a virtually unlimited food resource for pests thereby greatly increasing the threshold of density-dependent population re gulatory factors. An unusually dense population of prey enables an equally unusually dense population of predators to exist and therefore bottomup regulation across trophic levels, as opposed to top-down, is most likely to occur in vertebrate predator-prey systems in agricultural landscapes.

PAGE 40

40 Table 3-1. Life history parameters for cott on rats derived from Montague et al. 1990 and Whitaker and Hamilton 1998, and modified to fit abundance data collected in areas of the EAA that did not have Barn Owls. Parameter Value Inbreeding depression Recessive lethals model Breeding system Polygynous % males breeding 90% Breeding age for males & females 1 month Maximum age of reproduction 5 months Maximum # of broods/month 1 Mean # offspring/brood 6 Maximum # offspring/brood 8 Sex ratio at birth (% males) 50% Density dependent reproduction Yes % Females breeding at low density P(0) 75% or 45% % Females breeding at carrying capacity P(K) 50% or 45% Mortality from age 0 to 1 (SD) 45 (1) Annual mortality after age 1 (SD) 45 (1) Initial population size (with stable age distribution) 90 Carrying capacity (SD) 1750 (10)

PAGE 41

41 Table 3-2. Competing Huggins closed-capture models ( AICc < 2) for cotton rats, roof rats, and rice rats during early, mid, and late sugarcane growth phases. COTTON RAT Phase Model AICc AICc wi p(age) = c(age) 711.640.00 0.662 Early p(sex age) = c(sex age) 713.541.90 0.257 Mid p(sex*age) c(sex*age) 3893.730.00 0.624 p(sex*age) c(sex age) 11658.00.00 0.455 Late p(sex age) c(sex age) 11658.40.49 0.357 ROOF RAT Phase Model AICc AICc wi p(.) c(sex age) 818.550.00 0.188 p(.) c(age) 819.470.92 0.118 p(age) c(sex age) 819.661.11 0.107 Early p(sex*age) c(sex age) 820.451.90 0.072 p(.) c(sex*age) 1164.900.00 0.293 p(sex) c(sex*age) 1166.281.38 0.147 p(age) c(sex*age) 1166.351.45 0.142 Mid p(sex*age) c(sex*age) 1166.741.84 0.117 p(sex*age) c(sex) 2711.240.00 0.336 Late p(sex*age) c(sex age) 2712.541.30 0.175 RICE RAT Phase Model AICc AICc wi p(.) c(age) 1135.950.00 0.230 p(.) c(sex age) 1136.680.73 0.159 p(age) c(age) 1137.741.80 0.093 Early/Mid p(sex*age) c(age) 1137.911.97 0.086 p(.) c(age) 1318.920.00 0.279 p(sex) c(age) 1320.501.58 0.126 p(.) c(sex age) 1320.781.86 0.110 p(age) c(age) 1320.801.88 0.109 Late p(sex*age) c(age) 1320.871.95 0.105

PAGE 42

42 Table 3-3. Proc Mixed repeated measures ANOVA ex amining the influence of Barn Owl density and sugarcane growth on abundance of three species of rodents. COTTON RAT ROOF RAT RICE RAT Effect df F-value p-value F-value p-value F-value p-value Phase 253.06<0.0129.92<0.011.40 0.26 Year 1 0.05 0.83 0.74 0.400.26 0.62 Treatment 1 1.84 0.19 4.91 0.040.43 0.52 Treatment*Phase 2 1.27 0.30 4.16 0.030.96 0.39 Year*Phase 2 0.48 0.62 0.10 0.900.60 0.55 Year*Treatment 1 2.39 0.14 5.15 0.030.62 0.44 Year*Phase*Treatment 2 4.38 0.03 3.72 0.041.07 0.35 Table 3-4. Pair-wise least square mean comparisons among study site groupings for cotton rats and roof rats. Abundance of rodents in Experimental ’06 sites should have been less than that in all other sites if owls were biological controllers of rodents. COTTON RAT Comparison Expected Actual Control '05 vs. Control '06 = = (t=0.17, p=0.88) Experimental '05 vs. Experimental '06> = (t=2.37, p=0.10) Control '05 vs. Experimental '05 = = (t=0.69, p=0.54) Control '06 vs. Experimental '06 > = (t=1.50, p=0.23) ROOF RAT Comparison Expected Actual Control '05 vs. Control '06 = = (t=-2.18, p=0.12) Experimental '05 vs. Experimental '06> = (t=1.62, p=0.20) Control '05 vs. Experimental '05 = = (t=-2.47, p=0.09) Control '06 vs. Experimental '06 > = (t=0.76, p=0.50)

PAGE 43

43 Table 3-5. Results of Dunnett’s tests comparing abundance estimates of cotton rats derived from PVA model scenarios involving various dens ities of Barn Owls, and incorporating two harvesting intensities and two levels of cotton rat reproduction (A, 75% breeding; B, 45% breeding), to abundances derived from models with no owl predation. Simultaneous 95% confidence intervals that do not encompass zero indicate significance. A) 1rat/owl/month 3 rats/owl/month Nest boxes (Owls) Difference between means Simultaneous 95% confidence interval Difference between means Simultaneous 95% confidence interval 1 (4) -6.7-57.7, 44.2-18.1-156.9, 120.7 4 (16) -33.0-83.9, 17.9-126.8-265.6, 12.0 6 (24) -48.9-99.8, 2.1-344.1-482.9, -205.3 8 (32) -59.6-110.6, -8.6-686.9-825.7, -548.1 10 (40) -79.7-130.6, -28.7-1001.4-1140.1, -862.6 B) 1rat/owl/month 3 rats/owl/month Nest boxes (Owls) Difference between means Simultaneous 95% confidence interval Difference between means Simultaneous 95% confidence interval 1 (4) -7.4-142.6, 127.7-41.6-170.0, 86.8 4 (16) -199.5-334.6, -64.3-524.1-652.5, -395.7 6 (24) -242.2-377.4, -107.1-736.3-864.7, -607.9 8 (32) -265.0-400.2, -129.9-1054.5-1182.8, -926.1 10 (40) -462.5-597.7, -327.4-1206.4-1334.7, -1078.0

PAGE 44

44 Figure 3-1. Everglades Agricultural Area and locations of sites used for experimental manipulation (M1, M2, M3) and control (C1, C2, C3). Dots represent Barn Owl nest boxes installed prior to 2005, pluses represent boxes installed at the end of 2005 for this study. 5 km N Lake Okeechobee C 1 C 2 C 3 M 2 M 3 M 1

PAGE 45

45 Figure 3-2. Mean abundance estimates ( SE) for cotton rats, roof rats, and rice rats derived from Huggins closed capture modeling using Program MARK. Cotton Rats 0 20 40 60 80Control '05 Control '06 Experimental '05 Experimental '06 Roof Rats0 5 10 15 20Control '05Control '06Experimental '05Experimental '06Abundance (SE) Early Phase Mid Phase Late Phase Rice Rats0 5 10 15 20 Control '05Control '06Experimental '05Experimental '06

PAGE 46

46 0 250 500 750 1000 1250 1500 1750 2000 0146810 Number of occupied nest boxesCotton rat abundance (SD) 1 rat & 75% 3 rats & 75% 1 rat & 45% 3 rats & 45% Figure 3-3. Peak abundance estimates of cotton ra ts in a single sugarcane field derived from PVA model scenarios simulating various densities of occupied Barn Owl nest boxes with two levels of predation intensity (1 & 3 rats consumed per owl per month) and two levels of cotton rat reproductive rates (75% and 45% of females breeding). Asterisk below bar indicates a significant difference (p<0.05) from the reference estimate derived from the scenario with no nest boxes.

PAGE 47

47 CHAPTER 4 NEST BOXES, BARN OWLS, AND AGRICULTURE: DOES CROP HARVEST LEAD TO LOWERED REPRODUCTIVE SUCCESS? Introduction Avian species select habitat and nesting locations based on an assessment of habitat quality (Petit and Petit 1996; Muller et al. 1997; Martin 1998; Kolbe and Janzen 2002). Anthropogenic changes to habitat may cause environmental cues used by birds to assess quality to no longer provide an accurate gauge of habitat suitability (Misenhelter and Rotenberry 2000; Schlaepfer et al. 2002; Battin 2004; Weldon and Haddad 2005). An integral component of habitat suitability is temporal stability. Organisms must predict long-term suitability of habitat based on an immediate assessment of quality and do not have the ability to reject good habitat that will be altered in the future. Temporal instability in habitat quality may be especially consequential when selecting breeding sites (Orians and Wittenberger 1991). For example, grassland birds often establish nests in croplands because the structure of these systems mimics that of natural grassland habitat and because of the ample food supplies available in these areas. However, nests in croplands often are destroyed by harvesting activities (Best 1986; Bollinger et al. 1990). The decoupling of cue, quality, and temporal stability of habitat can lead to lowered reproductive success of individuals, complete failure of nests, and extirpation of populations (Kokko and Sutherland 2001; Shochat et al. 2005). Nest boxes are useful tools for expanding nest site availability for many cavity nesting bird species (e.g., Willner et al. 1983; Barry 1992). However, birds may be drawn to low quality sites by poorly placed boxes (e.g., Semel and Sherman 2001). Furthermore, boxes may be placed in areas that are considered by humans to be high quality based on an immediate assessment of surrounding habitat, but these sites may lack the temporal stability of resources

PAGE 48

48 required by the targeted species. Provision of artificial nesting structures may have negative consequences if the wider ecological context of this management action is not fully considered. Barn Owls (Tyto alba), which primarily are cavity nesters, occur in grasslands, agricultural fields, and other open habitat throughout most of the world (Marti 1992; Taylor 1994). This species is declining throughout North America because of habitat alteration and changes in agricultural practices (Taylor 1994). In the Everglades Agricultural Area (EAA) of south Florida, where sugarcane is the primary crop grown, abundance of Barn Owls historically has been limited because trees large enough to contain hollow spaces suitable for nesting are not common and buildings used by owls for nesting and roosting are disappearing rapidly because of hurricanes and general disuse. Over the past 20 years, approximately 200 Barn Owl nest boxes have been installed throughout the EAA to enhance the regional owl population with the intent of these raptors serving as biological controllers of pest rodents. At least two-thirds of the boxes are used by owls during the nesting season (late Aug – June), and all appear to have been used for nesting at least once (Martin, pers. obs.). Lo calized densities in the EAA can reach 18 pairs/5 km2 (Martin, pers. obs.), whereas densities in other parts of this species’ distribution range from 0.7 pairs/5 km2 to 2.6 pairs/5 km2 (Taylor et al. 1988; Altwegg et al. 2003). The objective of this study was to assess nesting success of the Barn Owl population in the EAA in relation to sugarcane harvest to determine if the temporal instability of the agricultural landscape may lead to lowered reproductive success. At the time when most owl nests are established, most of the sugarcane in EAA has not been harvested and rodents that occupy crop fields are a plentiful food resource fo r owls. However, nesting chronology of Barn Owls is relatively long (approximately 90 days; Marti 1992) and food availability declines precipitously following harvest at some point dur ing most nesting attempts (Chapter 3). I

PAGE 49

49 hypothesized that crop harvest surrounding occupied nest boxes would result in lowered nest survivorship. I predicted that in areas where sugarcane was harvested before chicks fledged, chicks would exhibit lower rates of growth than ch icks in areas that retained sugarcane. I also examined nest survivorship in relation to the proximity of major roadways because roadkill mortality of parental adults may serve as a secondary source of nest failure in the EAA (Martin, pers. obs.). Research Design I hypothesized that the daily survival rate of Barn Owl nests in the Everglades Agricultural Area would be lower after harvest of sugarcane. To test this hypothesis, I used a comparative observational approach. Active Barn Owl nest boxes were monitored throughout the nesting season and daily survival rates were calculated for preand post-harvest periods and nesting phases. I predicted that daily survival would be lower following harvest of all sugarcane fields adjacent to nest boxes and that the hatchling phase of nesting would have a lower daily survival rate than the incubation phase. I assumed that periodic observations of nests did not influence the fate of nests. Prior studies have indicated that the risk of observer-induced nest failure is minimal for this species, however I conducted nest visits at weekly intervals to further minimize this risk. Alternatively, daily survival of Barn Owl nests could have been influenced by date, age of nest, proximity of nests to well-traveled roadways, and climate (specifically temperature and rainfall). To test for these scenarios, I constructed a variety of predictive models of daily survival of Barn Owl nests and compared them using AIC. I also hypothesized that the daily survival rate of individual Barn Owl chicks would be lower when nests are surrounded entirely by harvested fields than when there is at least one field with standing sugarcane nearby due to the crash in rodent abundance that follows sugarcane harvest. To test this hypothesis, I used a comparative observation approach. I calculated daily

PAGE 50

50 survival for individual chicks from nests that either were or were not surrounded by standing sugarcane fields for the entire nesting period. I also measured and compared the growth rate of these chicks (mass, tarsus length, wing chord length). I predicted that the daily survival of chicks in nests surrounded by harvested fields would be lower than those in nests with at least one adjacent field with standing sugarcane. I also predicted that chicks that died before fledging in post-harvest nests would have a slower rate of growth than chicks in preor post-harvest nests that survived. I assumed that periodic observations of nests and handling of chicks did not influence fate of the chicks. I conducted nest visits and measured chicks at weekly intervals to minimize this risk. Alternatively, daily survival of Barn Owl chicks could have be influenced by hatch order or age of chick. To test for these scenarios, I constructed a variety of predictive models of daily survival of Barn Owl chicks and competed them using AIC. Methods Study location and species: The Everglades Agricultural Area is an approximately 160,000-ha region located in south Florida (Fig. 41). Sugarcane is the primary crop grown in this area. Individual sugarcane fields, measuring 200 x 800 m, are harvested from Nov Mar following a 10 14 month growing season. Harvest occurs in a patchwork manner, with only a few adjacent fields harvested simultaneously. As harvest progresses, the EAA landscape becomes an increasingly sparse patchwork of standing sugarcane. Rodents are abundant in sugarcane fields during most of the year (Holle r et al. 1981; Lefebvre et al. 1989; Martin et al. 2007). However, rodents do not remain in fi elds following harvest and regional rodent abundance likely drops greatly during this period resulting in an annual population cycle driven by the cropping cycle (Chapter 3). Barn Owls lay 4 – 7 eggs per nest (Marti 1992; Taylor 1994). Eggs are laid every 2 – 3 days until clutches are completed, but incubation begins after the first egg is laid. This results in

PAGE 51

51 asynchronous hatching, which begins approximately thirty days following the onset of incubation. A developmental lag between younger and older chicks within broods remains evident until fledging occurs approximately 60 days after hatching. If nesting attempts are completed or fail early in the season, second nesting attempts can occur. Monitoring overall nesting success: I monitored 93 active Barn Owl nest boxes in the EAA from 1 Dec 2004 – 31 Jan 2007, over 801 nest check intervals and 7231 exposure days. Most nests initiated before Nov 1 during the 2005 – 2006 nesting season were destroyed by Hurricane Wilma, therefore only nests initiated after this date were included from that nesting season. Approximately every 7 days, box contents were inspected visually or by using a video camera system mounted on the end of an extendable pole. If one egg was present following a check with no eggs, the nest was not checked for two weeks to minimize risk of nest abandonment. Following this, I visited active nests weekly until broods either fledged or failed and counted the number of eggs and/or chicks present during each visit. Nests were considered successful if at least two chicks survived to the expected fledging date for the nest, which was calculated as 60 days from the hatch date of the middle chick. I considered two chicks as the threshold for success, as opposed to one, because mortality for Barn Owls in their first year typically is high (Schifferli 1957; Frylestam 1972; Juillard and Beuret 1983). Number of days since the previous check and age of nest also were recorded during each visit. Mean temperature and total rainfall for each check interval were obtained from the Florida Automated Weather Network station at the Everglades Research and Education Center, Belle Glade, FL. Monitoring within-nest survivorship: In order to examine the relationship between sugarcane harvest and within-nest patterns of mortality, I monitored the fates of 84 chicks in a subset of 18 nests. Eleven “post-harvest” nests, monitored from 7 Jan – 23 May 2006, were

PAGE 52

52 selected in areas where all sugarcane fields within 1500 m of each nest box were harvested prior to nest initiation. This distance represents the approximate size of a Barn Owl home range (Hegdal and Blaskiewicz 1984; Marti 1992; Taylor 1994). Seven “pre-harvest” nests, monitored from 30 Aug – 22 Dec 2006, were selected in areas where harvesting activities in fields within 1500 m did not commence until after nests were completed. All nests were monitored following the previously described protocol; however, they were visited more frequently around the estimated hatch date (30 days from lay date; Marti 1992) in order to mark individual chicks. Colored nail polish was applied to talons of newly hatched chicks to facilitate identification of individuals. Once tarsi were fully developed (approximately 30 days), uniquely numbered USFWS lock-on leg bands were applied. After all chicks within a brood were marked, I checked nests weekly until they failed or fledged. During each visit, I recorded age, mass, tarsus length (tibio-tarsal articulation to palm), and wing chor d (wrist to tip of unflattened, closed wing) for each nestling. Habitat assessment: I assessed habitat around nests to examine the influence of sugarcane harvest on nest survivorship. I plotted the coordinates of all nest boxes onto a map of the EAA agricultural fields using ArcMap 9.2 (ESRI Inc., 2006). Crop type (sugarcane, rotational, or fallow) and harvest date of i ndividual sugarcane fields were obtained from landowners. I then created individual field maps for each year of the study that described landcover and harvesting chronology. Only fields with standing sugarcane were considered suitable sources of prey for owls. Rotational crops (e.g., sod, rice, sweet corn, leafy vegetables) and fallow fields were unlikely to harbor significan t rodent populations because of lack of cover, the presence of standing water, a short growing season with frequent soil disturbance, and/or frequent application of rodenticides. I determ ined if at least one standing sugarcane field

PAGE 53

53 occurred within an approximated home range distance of Barn Owls (1500-m radius; Marti 1992) around each box on each check date. Vehicle strikes may be a significant source of mortality for owls in the EAA (Martin, pers. obs.). Therefore, I also determined whether at least one major road occurred within 1500 m of each nest box using the 2000 TIGER Florida roads line file (US Census Bureau). Roads described in the dataset as “Primary” or “Secondary,” as well as those frequently used to access sugarcane mills, were considered major roads. Statistical analysis: I used the logistic-exposure method (Shaffer 2004) in Proc GENMOD (SAS 2003) to model the influence of various predictor variables on the daily survival rate of 93 nests. Prior to analysis, I developed the following set of candidate explanatory models: (1) DATE (date of check); (2) NESTAGE (age of nest); (3) TEMPERATURE and RAIN; (4) PHASE (incubation or hatchling); (5) SUGARCANE (presence/absence of standing sugarcane within 1500 m of nest); (5) ROAD (presence/absence of major road within 1500 m of nest); (6) SUGARCANE and ROAD; (7) PHASE and SUGARCANE; (8) PHASE and ROAD; and (9) PHASE, SUGARCANE, and ROAD. For the subset of nests in which fate of individual chicks was monitored, I modeled the daily survival rate for each chick during the nestling phase only by fitting logistic exposure models using Proc NLMIXED. Survival rates of chicks within nests are not independent from each other because all are exposed the same nest-wide stressors (e.g., poor parental condition). Therefore, NEST was included as a random variable in all models. I developed the following a priori predictive model set: (1) ORDER (order that chick hatched within brood); (2) AGE (age of chick); (3) HARVEST (preor post-harvest); (4) AGE and HARVEST; (5) ORDER and HARVEST; (6) ORDER and AGE; (7) HARVEST*AGE; (8) ORDER*AGE; (9) HARVEST

PAGE 54

54 and ORDER*AGE; and (10) ORDER and HARVEST*AGE. I was unable to include the interaction of HARVEST and ORDER in the model set because of insufficient data. Number of siblings was not included in the analysis because all brood sizes were similar ( x = 4.9, SE = 0.2). I verified the fit of the most parameterized model for each dataset using the HosmerLemeshow goodness-of-fit test (Hosmer and Lemeshow 2000). I used Akaike’s information criterion (Burnham and Anderson 2002) to rank competing models. Best models were selected by judging degree of support as measured by AICc and normalized Akaike weights. Models with AICc 2 were considered competitive. To account for uncertainty in model selection, I also calculated model-averaged weighted parameter estimates and their associated standard errors and 95% confidence intervals (Burnham and Anderson 2002, Shaffer 2004). I then used these values to calculate odds ratios and their associated 95% confidence intervals. Odds ratios provide a direct interpretation of the influence of a variable on nesting success (e.g., Peak et al. 2004, Lloyd and Martin 2005). Odds ratio values greater than one indicate a positive effect and those less than one indicate a negative effect. If the 95% CI of the ratio encompasses one, no significant effect is indicated. I estimated daily survival rates for each level of explanatory variables that had significant odds ratios using the range of observed values of the covariates, while holding all other covariates in the models constant (Shaffer and Thompson 2007). To examine the influence of sugarcane harvest on morphological development of chicks, I first divided the chicks into 4 groups: pre-harvest survived (n=23), pre-harvest died (n=5), post-harvest survived (n=28), and post-harvest died (n=28). Because sample size for pre-harvest chicks that died before fledging was small, I eliminated this group from the analysis. I then used a repeated measures approach in Proc Mixed (SAS) with NEST and AGE as random variables to compare mass, tarsus length, and wing chord among the three remaining groups. Appropriate

PAGE 55

55 covariance structures were selected using best fit statistics. Data from the two groups that survived to fledge were truncated to the maximum observed age of post-harvest chicks that died (54 days) to facilitate comparisons among groups. I used Bonferroni adjusted p-values to perform pair-wise comparisons among groups. This multiple-comparison correction is a more conservative approach than using standard p-values (Wright 1992). Results Twenty-two of the 93 nests (24%) failed. Eighty-two nests (88%) were established in nest boxes that were surrounded primarily by unharvested sugarcane fields at the time of nest initiation. Of these, sugarcane was removed entirely within 1500 m of 53 nests and was partially removed within 1500 m of 29 nests at some point during nesting. Eleven of the 93 nests (12%) were established in nest boxes after all sugarcane fields within 1500 m had been harvested. At least one major road occurred within 1500 m of fifty-eight nests (62%). Overall reproductive success: The most parameterized model adequately fit the observed values for overall nesting success ( 2 = 3.88, df = 6, P = 0.69). The model that contained independent effects of PHASE, ROAD, and SUGARCANE received the strongest support (Table 4-1). No other models were competitive. The confidence interval for the odds ratio of SUGARCANE indicated significance, and the odds ratio value suggested that the odds of a nest surviving from one day to the next was about 4 times higher when sugarcane was present within 1500 m of a nest than when it was absent (Table 4-2, Fig. 4-2A). There was a trend towards a lower daily survival rate during the hatchling phase than during the incubation phase, but this trend was not significant (Table 4-2, Fig. 4-2B). There also was a trend towards lower daily survival when major roads were near nests than when they were absent, but likewise this trend was not significant (Table 4-2, Fig. 4-2C).

PAGE 56

56 Within-nest survivorship: The most parameterized model adequately fit the observed values for within-nest survivorship ( 2 = 5.22, df = 7, P = 0.63). The model containing the effect of ORDER and the interaction of HARVEST a nd AGE received the strongest support (Table 43). No other models were competitive. Although the model averaged parameter estimate for ORDER was not significant (Table 4-4), this para meter was a significant factor (p < 0.05) in all individual models in which it was included. The interaction between HARVEST and AGE was significant (Table 4-4). Daily survival rate increased slightly with age in nests surrounded by sugarcane and decreased with age in nests surrounded by harvested fields (Fig. 4-3). Overall, daily survival appeared to be higher in nest s surrounded by standing sugarcane, but was highly variable in nests surrounded by harvested fields. Morphological development: Mass of individual chicks varied significantly depending on whether they were from nests prior to harvest and survived to fledge, were from post-harvest nests and survived, or were from post-harvest ne sts and died before fledging (F = 4.26, df = 2, p = 0.01; Fig. 4-4). Post-harvest chicks that di ed during nesting weighed significantly less than pre-harvest chicks and post-harvest chicks that survived to fledge (t = -2.56, pBonf = 0.03; t = 2.44, pBonf = 0.04, respectively). Mass did not differ between pre-harvest and post-harvest chicks that survived (t = -0.46, pBonf = 1.94). Tarsus length varied significantly among groups as well (F = 3.76, df = 2, p = 0.02), however this was not reflected in the Bonferroni pair-wise comparisons. Wing chord did not vary significantly among any groups (F = 1.24, df = 2, p = 0.29). Discussion Abundance of rodents in the EAA is linked to the annual cropping cycle of sugarcane (Chapter 3). Prior to harvest, rodents are extr emely abundant within individual fields. Rodent populations crash following the removal of food and cover resources via harvesting activities,

PAGE 57

57 and then increase throughout the following year as new sugarcane crops mature, and the food and cover that they provide for rodents is rest ored. Nesting success of some Barn Owls in the EAA clearly was negatively influenced by sugarcane harvest, likely because of harvest-induced crashes in their primary food source during nesting. Overall daily survival rate of owl nests wa s significantly lower after sugarcane adjacent to nests was removed. However, daily survival rates for nests were highly variable when sugarcane was absent. This variability may have been attributable to habitat characteristics not measured in this study, such as the presence of weedy ditch banks and refuse piles. These areas may serve as refugia habitats that allow rodents to persist locally following sugarcane harvest. Another potential source of this variability was variation in parental hunting ability, which may not have become apparent until food resources were relatively scarce in the landscape. The survivorship of individual chicks in nests surrounded by sugarcane was higher than that of chicks in nests where adjacent sugarcane had been completely harvested. Although monitoring of preand post-harvest nests was confounded with time, abundance of rodents in fields adjacent to nests was more likely to be associated specifically with the presence or absence of food and cover in those fields rather than with date in general. When sugarcane was absent and rodents likely were scarce in the vicinity of nests, daily survival rate of chicks decreased with age. Fo r birds in general, the likelihood of daily survival typically increases with nest age during the ne stling phase (e.g., Dinsmore et al. 2002; Grant et al. 2005). Nearly all of the chicks from nests surrounded by sugarcane for which I monitored morphological development survived to fledge. Conversely, half of those from nests where sugarcane had been harvested died during nesting. Chicks that died before fledging from nests where sugarcane had been harvested weighed less than all chicks that survived to fledge. This

PAGE 58

58 suggests that lower survivorship following harvest was attributable to diminished food resources during this period. Although some hatchling mortality may have been linked to parental mortality associated with roads, starvation likely was the primary source of hatchling mortality. A similar pattern of Barn Owl chick survivorship has previously been linked to vole cycling (~95% of chicks fledged during peak vole years, ~55% fledged in vole decline years; Taylor 1994). When most Barn Owls in the EAA select nesting locations, the entire landscape could be considered “good” because harvesting activities have not begun and therefore food resources are plentiful and widely distributed. The limiting factor of owl density at this time likely is nest site availability. The provision of nest boxes may enable the EAA owl population to exist at an extremely high density with breeding owls being very fecund if food availability remains stable over time. However, when harvesting begins and rodent populations decline, as occurs around almost all of the owl nests in the EAA to some degree, the system shifts towards an undesirable condition. By this point, owls are locked into nesting locations and they may experience an ecological trap that is temporal in nature. I did not collect long-term demographic data to test for population level effects, however this study shows that a mechanism for a trap exists in the form of temporal instability of habitat quality. Trap mechanisms that are spatial or behavioral in nature are well documented (Schlaepfer et al. 2002), however traps resulting from temporal shifts of habitat quality within breeding seasons are not. When evaluating how habitat variables influence nesting success of birds, and ultimately how these factors influence populations of birds as a whole, it is imperative to consider how the influence of habitat may shift over time. A highly productive population of Barn Owls in the EAA may contribute to the stability of the wider regional population by providing recr uits that disperse throughout the surrounding

PAGE 59

59 landscape. Young owls produced in the EAA may be capable of spreading throughout the southeastern United States into areas where the species is imperiled. Little is known about dispersal patterns of juvenile Barn Owls, but dispersal distances of up to 2000 km have been recorded (Stewart 1952, Soucy 1980). This topic merits further study to determine the extent to which metapopulation dynamics may exist for this species. Meanwhile, efforts should be made to maximize the reproductive output of Barn Owls in the EAA. This may be achieved by simply delaying harvest of sugarcane fields adjacent to active nest boxes until after nesting has been completed.

PAGE 60

60 Table 4-1. Model-fit for nest success analysis of 93 Barn Owl nests in the Everglades Agricultural Area, 2004-2007. Model K AICc AICc wi PHASE + ROAD + SURGARCANE 4179.13 0.00 0.747 ROAD + SURGARCANE 3182.79 3.66 0.120 PHASE + SURGARCANE 3183.37 4.25 0.089 PHASE + ROAD 3186.34 7.21 0.020 SURGARCANE 2186.44 7.31 0.019 DATE 2191.9012.78 0.001 PHASE 2192.0112.88 0.001 TEMPERATURE + RAIN 3192.6013.47 0.001 ROAD 2195.6616.53 0.000 NEST_AGE 2198.1018.97 0.000 CONSTANT 1200.5121.38 0.000 Table 4-2. Model-averaged parameter estimates and odds ratio values of predictor variables of Barn Owl nest success in the Everglades Agricultural Area, 2004-2007. Variable Model-averaged parameter estimate SE Odds ratio 95% CI SURGARCANE (Present vs. Absent) 1.460.564.29 1.42-13.01 PHASE (Incubation vs. Hatchling) 1.290.863.64 0.65-20.37 ROAD (Absent vs. Present) 1.190.703.29 0.81-13.40 DATE 0.000.001.00 1.00-1.00 NESTAGE 0.000.001.00 1.00-1.00 RAIN 0.000.001.00 1.00-1.00 TEMPERATURE 0.000.001.00 1.00-1.00

PAGE 61

61 Table 4-3. Model-fit for analysis of survival of 84 individual Barn Owl chicks in the Everglades Agricultural Area. Model K AICc AICc wi ORDER + HARVEST*AGE 6 186.73 0.00 0.990 HARVEST + ORDER*AGE 6 198.4811.74 0.003 ORDER + AGE + HARVEST 5 198.5511.81 0.003 ORDER + AGE 4 198.6411.91 0.003 ORDER*AGE 5 198.9012.17 0.002 HARVEST*AGE 5 212.8726.13 0.000 ORDER + HARVEST 4 214.1727.44 0.000 ORDER 3 214.8228.09 0.000 AGE + HARVEST 4 219.7733.04 0.000 AGE 3 221.2734.54 0.000 HARVEST 3 222.9036.17 0.000 CONSTANT 2 224.4537.72 0.000 Table 4-4. Model-averaged parameter estimates and odds ratio values for predictor variables of survival of 84 individual Barn Owl chicks in the Everglades Agricultural Area. Variable Model-averaged parameter estimate SE Odds ratio 95% CI ORDER -1.030.240.36 1.27-0.22 AGE -0.100.020.90 1.02-0.86 HARVEST (Postvs. Pre-harvest) -2.801.680.06 5.35-0.00 ORDER*AGE 0.000.001.00 1.00-1.00 HARVEST*AGE (Postvs. Pre-harvest) 0.120.041.13 1.04-1.05

PAGE 62

62 Figure 4-1. Location of the Everglades Agricultural Area. EVERGLADES AGRICULTURAL AREA Lake Okeechobee

PAGE 63

63 Figure 4-2. Daily survival rates ( 95% CI) for Barn Owl nests (A) when sugarcane was present and absent within 1500 m, (B) during incubation and hatchling nest phases, and (C) when roads were present or absent within 1500 m 0.988 0.991 0.994 0.997 1.000 EggHatchlingDaily Survival RateB) 0.988 0.991 0.994 0.997 1.000 No RoadsRoadsC) 0.988 0.991 0.994 0.997 1.000 Sugarcane PresentSugarcane AbsentA) 0.988 0.991 0.994 0.997 1.000 EggHatchlingDaily Survival RateB) 0.988 0.991 0.994 0.997 1.000 No RoadsRoadsC) 0.988 0.991 0.994 0.997 1.000 No RoadsRoadsC) 0.988 0.991 0.994 0.997 1.000 Sugarcane PresentSugarcane AbsentA)

PAGE 64

64 Figure 4-3. Daily survival rates ( 95% CI) of individual Barn Owl chicks by age of chick when sugarcane was present and absent. 0.64 0.70 0.76 0.82 0.88 0.94 1.00 010203040506070 Age (d)Daily Survival Rate Pre-harvest Post-harvest

PAGE 65

65 Figure 4-4. Growth curves depicting mean mass, tarsus length, and wing chord length ( SE) for Barn Owl nestlings. 20 30 40 50 60 70 80 010203040506070Tarsus (mm ) Post-harvest died Post-harvest survived Pre-harvest all 0 50 100 150 200 250 300 350 010203040506070 Age (d)Wing Cord (mm 100 200 300 400 500 600 700010203040506070Mass (g) survived) 20 30 40 50 60 70 80 010203040506070Tarsus (mm ) Post-harvest died Post-harvest survived Pre-harvest all 0 50 100 150 200 250 300 350 010203040506070 Age (d)Wing Cord (mm 100 200 300 400 500 600 700010203040506070Mass (g) survived)

PAGE 66

66 CHAPTER 5 CONCLUSION Barn Owls in the EAA prey heavily on cotton rats, roof rats, and rice rats. However, the results of this study do not indicate that Barn Owls are capable of reducing abundance of these crop pests. Reproductive capacity of all three species of rodent likely far exceeds rate of removal of individuals from the population via pr edation. Cotton rats are capable of producing litters of 10-15 young every month throughout the year, and roof rats and rice rats are only slightly less fecund (Whitaker and Hamilton 1998). Rodent populations occur at the scale of landscapes, but owl predation is a relatively localized phenomenon restricted primarily to areas near active nests for most of the year. Depredated rodents within foraging ranges of ow ls likely are replaced continually by dispersing individuals from the surrounding landscape. Incr easing density of Barn Owls across the EAA at the scale required to reduce regional rodent abundance is not feasible. The threshold of owl density required to produce a measureable decline in rodent populations is unobtainable high. Further studies are needed to determine if Barn Owl predation can reduce rodent abundance if combined with other pest control techniques, such as strategically timed application of rodenticides and elimination of rodent refugia habitat following harvest, in an integrated pest management approach. Abundance of rodents in the EAA is linked to the annual cropping cycle of sugarcane. Prior to harvest, rodents are extremely abundant within individual fields. Rodent populations crash following the removal of food and cover resources via harvesting activities, and then increase throughout the following year as new sugarcane crops mature, and the food and cover that they provide for rodents is restored. The survivorship of both nests and individual chicks

PAGE 67

67 was higher before sugarcane was harvested when compared to post-harvest conditions, likely due to harvest-induced crashes in rodents, the primary prey of the owls. In order for an ecological trap to exist, organisms must select poor-quality habitat over higher quality habitat and those organisms must suffer reduced fitness, potentially in the form of lowered reproductive output, as a result of this choice (Donovan and Thompson 2001; Robertson and Hutto 2006). In agricultural landscapes birds may not have a dichotomous choice of “good” vs. “bad” habitat because of the homogeneous nature of the landscape. For example, when most Barn Owls in the EAA select nesting locations, the entire landscape could be considered “good” because harvesting activities have not begun and therefore food resources are plentiful and widely distributed. The limiting factor of owl density at this time likely is nest site availability. The provision of nest boxes may enable the EAA owl population to exist at an extremely high density with breeding owls being very fecund if food availability remains stable over time. However, when harvesting begins and rodent popula tions decline, as occurs around almost all of the owl nests in the EAA to some degree, the system shifts towards an undesirable condition. By this point, owls are locked into nesting locations and they may experience an ecological trap that is temporal in nature. I did not collect long-term demographic data to test for population level effects of an ecological trap, however this study s hows that a mechanism for a trap exists in the form of temporal instability of habitat quality. Trap mechanisms that are spatial or behavioral in nature are well documented (Schlaepfer et al. 2002), however traps resulting from temporal shifts of habitat quality within breeding seasons are not. When evaluating how habitat variables influence nesting success of birds, and ultimately how these factors influence populations of birds as a whole, it is imperative to consider how the influence of habitat may shift over time.

PAGE 68

68 Although augmentation of owl density through installation of nest boxes does not appear to be an effective means of rodent control in large agricultural systems, this management tool may be important for conservation of owls as traditional nest sites of this species become scarce. As in previous studies (e.g., Muoz and Mur a 1990; Lee 1997; Moore et al. 1998; Potter 2004), I found that Barn Owl density can easily be manipulated by installing nest boxes. However, the success of this management technique exhibited in this study may have been greater than that in other areas because of extremely high localized abundance of owls in the EAA. Nevertheless, although Barn Owl management programs involvi ng nest boxes should not expect colonization to the degree achieved in the EAA, nest boxes are a proven management option where increasing abundance of the species is desirable. A highly productive population of Barn Owls in the EAA may contribute to the stability of the wider regional population by providing recr uits that disperse throughout the surrounding landscape. Young owls produced in the EAA may be capable of spreading throughout the southeastern United States into areas where the species is imperiled. Little is known about dispersal patterns of juvenile Barn Owls, but di spersal distances of approximately 2000 km have been recorded (Soucy 1980). This topic merits further study to determine the extent to which metapopulation dynamics may exist for this species. Meanwhile, efforts should be made to maximize the reproductive output of Barn Owls in the EAA. This may be achieved by simply delaying harvest of sugarcane fields adjacent to active nest boxes until after nesting has been completed.

PAGE 69

69 LIST OF REFERENCES Abarca, M.R. 1981. Rodent control in sugarcane fields. In: Proc. 2nd International-American Sugar Cane Seminar – Insect and Rodent Pests. 6-8 October 1981, Miami, FL, pp. 189200. Altwegg, R., A. Roulin, M. Kestenholz, and L. Jenni. 2003. Variation and covariation in survival, dispersal, and population size in barn owls Tyto alba. Journal of Animal Ecology 72:391-399. Bailey, J.A. 1984. Principles of Wildlife Management. John Wiley & Sons, New York. Barry, S.C. 1992. Evaluation of a Wood Duck nest box program at Yazoo National Wildlife Refuge, Mississippi, 1987-91. M.S. thes is, Mississippi State University. Battin, J. 2004. When good animals love bad habitats: ecological traps and the conservation of animal populations. Conservation Biology 18(6):1482-1491. Baucum, L.E., R.W. Rice, and T. J. Schueneman. 2006. An overview of Florida sugarcane. University of Florida, Institute of Food and Agricultural Science Extension. Accessed January 30, 2007 from http://edis.ifas.ufl.edu/SC032 Bayliss, P. and D. Choquenot. 2002. The numerical response: rate of increase and food limitation in herbivores and predators. Philosophical Tr ansactions of the Royal Society of London B 357:1233-1248. Best, L.B. 1986. Conservation tillage: ecological traps for nesting birds? Wildlife Society Bulletin 14:308-317. 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. Bonal, R. and J.M. Apricio. 2008. Evidence of prey depletion around lesser kestrel (Falco naumanni) colonies and its short term negative consequences. Journal of Avian Biology 39(2):189-197. Borkhataria, R.R., J.A. Collazo, and M.J. Groom 2006 Additive effects of vertebrate predators on insects in a Puerto Rican coffee planta tion. Ecological Applications 16(2): 696-703. Caltagirone, L.E. 1981. Landmark examples in cla ssical biological control. Annual Review of Entomology 26:213-232. Camerson, G.N. and S.R. Spencer. 1985. Assessment of space-use patterns in the hispid cotton rat. Oecologia 68(1):133-139. Conn, P.B., A.D. Arthur, L.L. Bailey, and G. R. Singleton. 2006. Estimating the abundance of mouse populations of known size: promises and pitfalls of new methods. Ecological Applications 16(2):829-837.

PAGE 70

70 Cooch, E. and G. White. 20078. Program MARK: A gentle introduction. Seventh Edition. URL:http://www.phidot.org/software/mark/docs/book/ Davis, D.E., K. Myers and J.B. Hoy. 1976. Biol ogical control among vertebrates. Pages 501-519 in C.B. Huffaker and P.S. Messenger, eds. Theory and Practice of Biological Control. Academic Press, New York. de Azevedo, F.C.C. and D.L. Murray. 2007. Spatial organization and food habits of jaguars (Panthera onca) in a floodplain forest. Biol ogical Conservation 137(3):391-402. Dempster, J.P. 1983. The natural control of populations of butterflies and moths. Biological Reviews 58:461-481. Deren, C.W., B. Glaz, P.Y.P. Tai, J.D. Miller, and J.M. Shine. 1991. Registration of ‘CP 801743’ sugarcane. Crop Science 31:235-236. Diffendorfer, J.E. and N.A. Slade. 2002. Long-distance movements in cotton rats (Sigmodon hispidus) and prairie voles (Microtus ochrogaster) in northeastern Kansas. American Midland Naturalist 148:309-319. Dinsmore, S.J., G.C. White, and F.L. Knopf. 2002. Advanced techniques for modeling avian nest survival. Ecology 83(12):3476-3488. Dobson, P. and D. Wexlar. 1979. Taphonomic inves tigations of owl pellets. Paleobiology 5:275284. Donovan, T.M. and F.R. Thompson. 2001. Modeling the ecological trap hypothesis: a habitat and demographic analysis for migrant s ongbirds. Ecological Applications 11:871-882. Doty, J.W. 1960. Rodent control in commercial cane areas of Florida Everglades. Proceedings of the ISSCT 10:1011-1015. Duckett, J.E. 1991. Management of the Barn Owl (Tyto alba javanica) as a predator of rats in oil palm (Elaeis quineensis) plantations in Malaysia. Birds of Prey Bulletin 4:11-23. Errington, P.L. 1932. Food habits of southern Wi sconsin raptors. Part 1. Owls. Condor 34:176180. Errington, P.L. 1946. Predation and vertebrate populations. The Quarterly Review of Biology 21:144-177, 221-245. Errington, P.L. 1967. Of Predation and Life. Iowa State University Press, Ames. Fagerstone, K.A., L.A. Miller, K.S. Bynum, J.D. Eisemann and C. Yoder. 2006. When, where and for what wildlife species will contraception be a useful management approach? Proceedings of the 22nd Vertebrate Pest Conference, p. 45-54.

PAGE 71

71 Forys, E.A. and R.D. Dueser. 1993. Inter-island movements of rice rats (Oryzomys palustris). American Midland Naturalist 130:408-412. Frylestam, B. 1972. Uber Wanderungen und St erblichkeit beringter skandinavischer Schleiereulen Tyto alba. Ornis Scandinavia 3:45-54. Garlough, F.E. 1938. Rodents in relation to sugar cane growing. Proceedings of the ISSCT 6:98106. Geiser, F., W. Westman, B.M. McAllan, and R.M. Brigham. 2006. Development of thermoregulation and torpor in a marsupial: energetic and evolutionary implications. Journal of Comparative Physiology B Biochemical Systemic and Environmental Physiology 176(2): 107-116. Grant, T.A., T.L. Shaffer, E.M. Madden and P.J. Pietz. 2005. Time-specific variation in passerine nest survival: New insights into old questions. The Auk 122(2):661-672. Guynn, D.C. 1993. Contraception in wildlife management: reality or illusion? USDA National Wildlife Research Center Symposia: Contraception in Wildlife Management, p. 241-246 Hagen, K.S., S. Bombosch, J.A. McMurtry. 1976. The biology and impact of predators. Pages 93-142 in C.B. Huffaker and P.S. Messenger, eds. Theory and Practice of Biological Control. Academic Press, New York. Hall, T.R., W.E. Howard and R.E. Marsh. 1981. Raptor use of artificial perches. Wildlife Society Bulletin 9(4):296-298. Hassell, M.P. 1985. Insect natural enemies as regulating factors. Journal of Animal Ecology 54:323-334. Hegdal, P.L. and R.W. Blaskiewicz. 1984. Evaluation of the potential hazard to Barn Owls of talon (brodifacoum bait) used to control rats and house mice. Environmental Toxicology and Chemistry 3:167-179. Holler, N.R., L.W. Lefebvre, and D.G. Decker. 1981. Ecology and control of rodent depredations to Florida sugarcane. Proceedings of the ASSCT 10:183-188. Holler, N.R., L.W. Lefebvre, and D.G. Decker. 1982. Ecology and control of rodent depredations to Florida sugarcane. In: Proc. 2nd Inter-American Sugar Cane Seminar – Insect and Rodent Pests. 6-8 October 1981, Miami, FL, pp. 183-188. Horn, D.J. 1988. Ecological approach to pest management. Guilford Press, New York, NY. Hoskins, W.M., A.D. Borden and A.E. Michelbacher. 1939. Recommendations for a more discriminating use of insecticides. Proceedings of the 6th Pacific Science Congress 5:199123.

PAGE 72

72 Howard, W.E. 1967. Biological control of ve rtebrate pests. Proceedings of the 3rd Vertebrate Pest Conference, pp. 137-157 Juliard, M. and J. Beuret. 1983. L’amenagement de sites de nidification et son influence sur une population de Chouettes effraies, Tyto alba, dans le nord-ouest de la Suisse. Nos Oiseaux 37:1-20. Kale, H.W. 1968. The relationship of Purple Martins to mosquito control. Auk 85:654-661. Kay, B.J., L.E. Twig, T.J. Korn, and H.I. Nicol. 1994. The use of artificial perches to increase predation on House Mice (Mus domesticus) by raptors. Wildlife Research 21:95-106. Kogan, M. 1998. Integrated pest management: historical perspectives and contemporary developments. Annual Review of Entomology 43:243-270. Kokko, H. and W.J. Sutherland. 2001. Ecological traps in changing environments: ecological and evolutionary consequences of a behaviour ally mediated Allee effect. Evolutionary Ecology Research 3:537-551. Kolbe, J.J. and F.J. Janzen. 2002. Impact of nest-site selection on nest success and nest temperature in natural and disturbed habitats. Ecology 83(1):269-281. Korpimaki, E. and K. Norrdahl. 1989. Predation of Tengmalm’s owls: numerical responses, functional responses and dampening impact on population fluctuations of microtines. Oikos 54:154-164. Krebs, C.J. and J.H Myers. 1974. Population cycles in small mammals. Advances in Ecological Research 8:267-399. Lacy, R.C., M. Borbat, and J.P. Pollack. 2008. Vortex: A stochastic simulation of the extinction process. Version 9.94. Chicago Zoologi cal Society, Brookfield, Illinois. Lee, C.H. 1997. Barn Owl for field rat control in cocoa. Journal of Tropical Agriculture and Food Science 25(1):43-54. Lefebvre, L.W., R.M. Engeman, D.G. Decker, and N.R. Holler. 1989. Relationship of roof rat population indices with damage to sugarcane. Wildlife Society Bulletin 17:41-45. Lefebvre, L.W., N.R. Holler, and D.G. Decker. 1985. Efficacy of aerial application of a 2% zinc phosphide bait on roof rats in sugarcane. Wildlife Society Bulletin 13:324-327. Lefebvre, L.W., C.R. Ingram and M.C. Yang. 1978. Assessment of rat damage to Florida sugar cane in 1975. Proceedings of the ASSCT 7:75-80. Lefebvre, L.W., D.L. Otis, and N.R. Holler. 1982. Comparison of open and closed models for cotton rat population estimates. Journal of Wildlife Management 46(1):156-163.

PAGE 73

73 Lenton, G.M. 1980. The ecology of Barn Owls (Tyto alba) in the Malay Peninsular with reference to their use in rodent control. Ph.D. thesis, University of Malaya, Kuala Lumpur Lindzey, F.G., V. Sickle, B.B. Ackerman, D. Barnhurst, T.P. Hemker, and S.P. Laing. 1994. Cougar population dynamics in southern Utah. Journal of Wildlife Management 58:619624. MacArthur, R. 1955. Fluctuations of animal populations, and a measure of community stability. Ecology 36:533-536. Marsh, R.E. 1998. Barn Owl nest boxes offer no solution to pocket gopher damage. Proceedings of the 18th Vertebrate Pest Conference 414-415. Marti, C.D. 1992. Barn Owl. Pp. 5-15 in The Birds of North America – Life Histories for the 21st Century (A.F. Poole, P. Stettenheim and F.B. Gill, eds.). The American Ornithologists’ Union and The Academy of Natural Sciences of Philadelphia, Washington, D.C. and Philadelphia. Martin, J.M., R.N. Raid, and L.C. Branch. 2007. Quantifying damage potential of three rodent species on sugarcane. Journal of the Amer ican Society of Sugarcane Technologists 27:4854. Martin, T.E. 1998. Are microhabitat preferences of coexisting species under selection and adaptive? Ecology 79:656-670. Meserve, P.L., J.R. Gutirrez, and F.M. Jaksic. 1993. Effects of vertebrate predation on a caviomorph rodent, the degu (Octodon degus), in a semiarid thorny scrub community in Chile. Oecologia 94:153-158. Misenhelter, M.D. and J.T. Rotenberry. 2000. Choices and consequences of habitat occupancy and nest site selection in Sage Sparrows. Ecology 81(10):2892-2901. Montague, C.L., L.W. Lefebvre, D.G. Decker, and N.R. Holler. 1990. Simulation of cotton rat population dynamics and response to rodenticide applications in Florida sugarcane. Ecological Modelling 50:177-203. Moore, T., D. Van Vuren and C. Ingels. 1998. Ar e Barn Owls a biological control for gophers? Evaluating effectiveness in vineyards and orchards. Proceedings Vertebrate Pest Conference 18:394-396. Muller, K.L., J.A. Stamps, V.V. Krishnan, and N.H. Willits. 1997. The effects of conspecific attraction and habitat quality on habitat selection in territorial birds (Troglodytes aedon). American Naturalist 150:650-661. Mumford, J.D. and G.A. Norton. 1984. Economics of decision making in pest management. Annual Review of Entomology 29:157-174.

PAGE 74

74 Muoz, A. and R. Mura. 1990. Control of small mammals in a pine plantation (central Chile) by modification of the habitat of predator s (Tyto alba and Pseudalopex sp.). Acta Oecologia 11:251-261. Nettles, V.F. 1997. Potential consequences and problems with wildlife contraceptives. Reproductive Fertility and Development 9:137-143. Newsome, A.E. 1990. The control of vertebrate pe sts by vertebrate predators. Trends in Ecology and Evolution 5:187-191. Newton, I. and M. Marquiss. 1991. Removal experiments and the limitation of breeding density in Sparrowhawks. Journal of Animal Ecology 60:535-544. Norrdahl, K. and E. Korpimaki. 1995. Effects of predator removal on vertebrate prey populations. Journal of Mammalogy 45:241-248. Orians, G.H. and J.F. Wittenberger. 1991. Spatial and temporal scales in habitat selection. The American Naturalist 137:S29-S49. Pearson, O.P. 1966. The prey of carnivores during one cycle of mouse abundance. The Journal of Animal Ecology 35:217-233. Petit, L.J. and D.R. Petit. 1996. Factors governing habitat selection by Prothonotary Warblers: field tests of the Fretwell-Lucas models. Ecological Monographs 66:367-387. Potter, L. 2004. Latitudinal life-history patterns of the Barn Owl (Tyto alba), and its potential role as an agent of biocontrol. MSc Thesis University of Cape Town, South Africa. Potts, G.R. 1980. The effects of modern agriculture, nest predation and game management on the population ecology of Partridges Perdix perdix and Alectoris rufa. Advances in Ecological Research 11:2-79. Pryde, M., P. Dilks, and I. Fraser. 2005. The home range of ship rats (Rattus rattus) in beech forest in the Eglinton Valley, Fiordland, New Zealand: a pilot study. New Zealand Journal of Zoology 32:139-142. Quinn, J.L. and W. Cresswell. 2004. Predator hun ting behaviour and prey vulnerability. Journal of Animal Ecology 73:143-154. Raczynski, J. and A.L. Ruprecht. 1974. The effect of digestion on the osteological composition of the owl pellets. Acta Ornithologica 15:1-12 Regisser, B. 1991. Bilan de 13 annes de protection et de suivi de la Chouette effraie, Tyto alba, dans le Haut-Rhin (France) de 1978 1990. Rapace Nocturnes, Actes du 30 Colloque interrgional d’ornithologie Porrentruy (Suisse). Reinert, S.E. 1984. Use of introduced perches by ra ptors: experimental results and management implications. Raptor Science 18(1):25-29.

PAGE 75

75 Rice, M.E. and G.E. Wilde. 1988. Experimental evaluation of predators and parasitoids in suppressing greenbugs (Homoptera: Aphididae) in sorghum and wheat. Environmental Entomology 17:836-841. Robertson, B.A. and R.L. Hutto. 2006. A framew ork for understanding ecological traps and an evaluation of existing evidence. Ecology 87(5):1075-1085. Samol, H.H. 1972a. Rat damage to sugarcane in Florida. Proceedings of the International Society of Sugarcane Technologists 14:575-580. Samol, H.H. 1972b. Rat damage and control in the Florida sugarcane industry. Proc. ISSCT 14:503-506. Sauer, J.R. and N.A. Slade. 1985. Mass-based demography of a Hispid Cotton Rat (Sigmodon hispidus) population. Journal of Mammalogy 66(2):316-328. SAS. 2003. SAS STAT Users Guide, Version 9.1, SAS Institute Inc., Cary, NC, USA. Schifferli, A. 1957. Alter und Sterblichkeit bei Waldkauz (Strix aluco) und Schleiereule (Tyto alba) in der Schweiz. Ornith Beob. 54:50-56. Schlaepfer, M.A., M.C. Runge, and P.W. Sher man. 2002. Ecological and evolutionary traps. Trends in Ecology & Evolution 17(10):474-480. Semel, B. and P.W. Sherman. 2001. Intraspecific parasitism and nest-site competition in wood ducks. Animal Behaviour 61(4):787-803. Shelford, V.E. 1942. Biological control of rode nts and predators. Science Monthly 55(4):331341. Shochat, E., M.A. Patten, D.W. Morris, D.L. Reinking, D.H. Wolfe, and S.K. Sherrod. 2005. Ecological traps in isodars: effects of tallgrass prairie management on bird nest success. Oikos 111:159-169. Sinclair, A.R.E., P.D. Olsen and T.D. Re dhead. 1990. Can predators regulate small mammal populations? Evidence from house mouse outbreaks in Australia. Oikos 59:382-392. Slade, N.A. and R.K. Swihart. 1983. Home range indices for the hispid cotton rat (Sigmodon hispidus) in northeastern Kansas. Journal of Mammalogy 64:580-590. Smith, R.F. 1978. History and complexity of integrated pest management. In Pest Control Strategies, ed. E.H. Smith, D. Pimentel, p. 41-53. New York: Academic. Solomon, M.E. 1949. The natural control of animal populations. Journal of Animal Ecology 18:1-35. Stewart, P.A. 1952. Breeding behavior and longev ity of banded Barn Owls in North America. The Auk 69(3):227-245.

PAGE 76

76 Sulok, M., N.A. Slade, and T.J. Doonan. 2004. Effects of supplemental food on movements of cotton rats (Sigmodon hispidus) in northeastern Kansas. Journal of Mammalogy 85(6):1102-1105. Sweetman, H.L. 1958. The Principles of Biological Control. Brown, Dubuqe, IA. Symondson, W.O.C., K.D. Sunderland and M.H. Greenstone. 2002. Can generalist predators be effective biocontrol agents? Annual Review of Entomology 47:561-594. Tamarin, R.H. and S.R. Malecha. 1972. Reproductive parameters in Rattus rattus and Rattus exulans of Hawaii, 1968 to 1970. Journal of Mammalogy 53(3):513-528. Taylor, I. 1994. Barn Owls: predator-prey relati onships and conservation. University Press, Cambridge. Taylor, I.R., A. Dowell, T. Irving, I.K. La ngford, and G. Shaw. 1988. The distribution and abundance of the Barn Owl Tyto alba in south-west Scotland. Scotish Birds 15:40-43. Votier, S.C., S. Bearhop, J.E. Crane, J.M. Arcos, and R.W. Furness. 2007. Seabird predation by great skuas Stercorarius skua – intra-specific competition for food? Journal of Avian Biology 38(2):234-246. Walsh, L.E., N.R. Holler, D.G. Decker, and C. R. Ingram. 1976. Studies of rodent damage and rodent population dynamics in Florida suga rcane. Proceedings of the ASSCT 5:227-230. Warren, R.J. 1995. Should wildlife biologists be invol ved in wildlife contraception research and management? Wildlife Society Bulletin 23(3):441-444. Weldon, A.J. and N.M. Haddad. 2005. The effects of patch shape on Indigo Buntings: evidence for an ecological trap. Ecology 86(6):1422-1431. Whitaker, J.O. and W.J. Hamilton. 1998. Mammals of the eastern United States, 3rd edition. Cornell University Press, Ithaca, NY. 583 p. White, G.C. 2002. Discussion comments on: the use of auxiliary variables in capture-recapture modeling. An overview. Journal of Applied Statistics 29(1-4):103-106. White, G.C. and K. P. Burnham. 1999. Program MARK: Survival estimation from populations of marked animals. Bird Study 46 Supplement: 120-138. Wilner, G.R., J.E. Gates, and W.J. Devlin. 1983. Nest box use by cavity-nesting birds. American Midland Naturalist 109(1):194-201. Wright, S.P. 1992 Adjusted P-values and simultaneous inference. Biometrics 48: 1005-1013. Ziesemer, F. 1980. Siedlungsdichte und Bruterfo lg von Schleiereulen Tyto alba in einer Probeflche vor und nach dem Anbringe n von Nisthilfen. Vogelwelt 101:61-66.

PAGE 77

77 BIOGRAPHICAL SKETCH Jason Martin received a bachelor’s degree in biology from Ramapo College of New Jersey in 1998. In fall 2001, he entered the graduate program at the University of Illinois in Urbana–Champaign and earned a Master of Science degree in natural resources and environmental science in 2003. In fall 2003, he enrolled in the graduate program at the University of Florida and earned a PhD in wildlife ecology and conservation in 2009.