<%BANNER%>

Ecological and Human Dimensions of the Monk Parakeet Damage to Crops in Argentina

Permanent Link: http://ufdc.ufl.edu/UFE0043559/00001

Material Information

Title: Ecological and Human Dimensions of the Monk Parakeet Damage to Crops in Argentina
Physical Description: 1 online resource (156 p.)
Language: english
Creator: Canavelli, Sonia B
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2011

Subjects

Subjects / Keywords: abundance -- argentina -- corn -- damage -- demographic -- farmer -- nest -- parakeet -- preferences -- psychological -- socio -- sunflower
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: The Monk Parakeet (Myiopsitta monachus) is considered among the most important bird pest species causing damage to crops in Argentina. In this study, I explored habitat features influencing abundance and damage of monk parakeets to crop fields and density of nests in inhabited farms with eucalyptus trees at multiple spatial levels. Additionally, I examined socio-psychological and socio-demographic factors influencing farmers´ preferences about management of monk parakeet damage to crops. Monk parakeet abundance and damage was greater in sunflower than in corn fields. Landscape variables, such as distance to nearest sites with trees, percentage of landscape with trees, and availability of foraging sites for monk parakeets around the crop fields, were more important than local variables in explaining monk parakeet damage to crop fields. However, local variables, such as field area, plant density and percentage of field border with trees, also were related to damage. Conversely, the density of monk parakeet nests in inhabited farms with eucalyptus trees was not clearly explained by any variable or combination of variables modeled in this study. Farmers preferred population control strategies, such as nest destruction and killing of birds, for decreasing monk parakeet damage to crops. Preferences of farmers for management strategies were related more strongly to attitudes toward monk parakeets than to any other factor considered in this study. Other important socio-psychological factors were perceived efficacy and previous knowledge about management strategies. Perceptions of magnitude of damage by monk parakeets practically were not related to preferences. Socio-demographic factors, such as age and education, were related to preferences in different ways depending on the management strategy. Based on this study, managers should consider both local and landscape factors when planning management measures to prevent monk parakeet damage to crop and reduce nesting on farms. Additionally, extension actions should be oriented to modifying attitudes toward monk parakeets as well as communicating and showing the efficacy of alternative management strategies. Given the current uncertainties in the outcome of management actions, an adaptive management approach would be useful to evaluate the efficacy of strategies other than lethal or reproductive control.
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 Sonia B Canavelli.
Thesis: Thesis (Ph.D.)--University of Florida, 2011.
Local: Adviser: Branch, Lyn C.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2014-12-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2011
System ID: UFE0043559:00001

Permanent Link: http://ufdc.ufl.edu/UFE0043559/00001

Material Information

Title: Ecological and Human Dimensions of the Monk Parakeet Damage to Crops in Argentina
Physical Description: 1 online resource (156 p.)
Language: english
Creator: Canavelli, Sonia B
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2011

Subjects

Subjects / Keywords: abundance -- argentina -- corn -- damage -- demographic -- farmer -- nest -- parakeet -- preferences -- psychological -- socio -- sunflower
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: The Monk Parakeet (Myiopsitta monachus) is considered among the most important bird pest species causing damage to crops in Argentina. In this study, I explored habitat features influencing abundance and damage of monk parakeets to crop fields and density of nests in inhabited farms with eucalyptus trees at multiple spatial levels. Additionally, I examined socio-psychological and socio-demographic factors influencing farmers´ preferences about management of monk parakeet damage to crops. Monk parakeet abundance and damage was greater in sunflower than in corn fields. Landscape variables, such as distance to nearest sites with trees, percentage of landscape with trees, and availability of foraging sites for monk parakeets around the crop fields, were more important than local variables in explaining monk parakeet damage to crop fields. However, local variables, such as field area, plant density and percentage of field border with trees, also were related to damage. Conversely, the density of monk parakeet nests in inhabited farms with eucalyptus trees was not clearly explained by any variable or combination of variables modeled in this study. Farmers preferred population control strategies, such as nest destruction and killing of birds, for decreasing monk parakeet damage to crops. Preferences of farmers for management strategies were related more strongly to attitudes toward monk parakeets than to any other factor considered in this study. Other important socio-psychological factors were perceived efficacy and previous knowledge about management strategies. Perceptions of magnitude of damage by monk parakeets practically were not related to preferences. Socio-demographic factors, such as age and education, were related to preferences in different ways depending on the management strategy. Based on this study, managers should consider both local and landscape factors when planning management measures to prevent monk parakeet damage to crop and reduce nesting on farms. Additionally, extension actions should be oriented to modifying attitudes toward monk parakeets as well as communicating and showing the efficacy of alternative management strategies. Given the current uncertainties in the outcome of management actions, an adaptive management approach would be useful to evaluate the efficacy of strategies other than lethal or reproductive control.
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 Sonia B Canavelli.
Thesis: Thesis (Ph.D.)--University of Florida, 2011.
Local: Adviser: Branch, Lyn C.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2014-12-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2011
System ID: UFE0043559:00001


This item has the following downloads:


Full Text

PAGE 1

1 ECOLOGICAL AND HUMAN DIMENSIONS OF THE MONK PARAKEET DAMAGE TO CROPS IN ARGENTINA By SONIA BEATRIZ CANAVELLI A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUI REMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2011

PAGE 2

2 2011 Sonia Beatriz Canavelli

PAGE 3

3 To my husband, Carlos, and our lovely daughters, Luciana and Eugenia To my parents, Marina ( deceased) and Juan Carlos

PAGE 4

4 ACKNOWLEDGMENT S I sincerely thank my advisor, Lyn Branch, for her continuous support and a ssistance on all stages of my dissertation. I greatly appreciate t he time and effort she dedicated to my professional development and this dissertation. I also thank my graduate committee, Marylyn Swisher, Katie Sieving Michael Avery and Peter Frederick, for their advice and continuous stimulus to conduct and finish this work. This project would have not been possible without the help of many people at the Institute of Agriculture Technology (INTA) in Argentina. Mara Elena Zaccagnini gave me not only her continuous support, but also the opportunity to conduct this project as part of an institutional project. Noelia Calamari helped me with the analyses of satellite images, statisti cal analyses with GLM, and methodological issues. Laura Addy Orduna assisted me with field work and also with research activities complementary to this project. Gregorio Gavier helped me by discussing methodological aspects of the research and Cristina Gonzlez helped me with preliminary statistical analyses. The extension agents Ricardo De Carli and Mara Jos Marnetto helped me with farmers interviews. Sebastin Dardanelli Guillermo Stamatti and Ricardo De Carli assisted with complementary institutional activities during the last p art of the dissertation. Mabel Rodriguez and Patricia Engl er helped me with socioeconomic infor mation about farmers in Entre R os. Marcelo Wilson, Dante Bedendo, Gloria Pausich, Adriana Saluso, Hugo Peltzer, Octavio Caviglia, Oscar Valentinuz assisted with satellite image or agronomic information pertaining to this project. Finally, Norma Formento, Hugo Tassi and O svaldo Paparotti gave me the institutional support and stimulus to pursue and finish this project Thanks to all o f them.

PAGE 5

5 I am also extremely grateful to my field assistant and colleagues Pedro Cavallero and Ricardo Jurez and my unconditional field assistant Benito Jauberts. I also thank Florencia Barcarolo, who helped me with the personal interviews of farmers and to m any land owners in Entre Ros, who kindly provided access to their farms This work would not have been possible without their collaboration. Additionally, I would like to thank colleagues from other institutions, who helped me with methodological asp ects of the research, particularly to Miller Curtis and James Colee (University of Florida, IFAS Statistical Consultant), Stella Vaira, Elena Carrera, Paula Ricardi and Mauro Alexis (Universidad Nacional del Litoral, Santa Fe), Cecilia Bruno (U niversidad N acional de Crdoba) and Silvana Sione (Univ ersidad Nacional de Entre Ros). Finally, and above all, I would like to thank my husband, Carlos Cappellacci and our daughters, Luciana and Eugenia; my parents, Marina Judith Gariboldi (deceased) and Juan Carlos Canavelli, my sister and brother, Ana and Alberto Canavelli, and their respective families, and all my friends, for their love, patience and continuous support during this process. Thanks also to Miri am BiaseCabrol for taking care of the house and the gir ls while I was at work. The National Institute of Agricultural Technology (INTA, Argentina) the University of Florida (UF) and the Autonomous University of Entre Ros (UADER) gave financial assistance for th e Ph.D. program and this project. Thanks to all of them.

PAGE 6

6 TABLE OF CONTENTS page ACKNOWLEDGMENTS .................................................................................................. 4 LIST OF TABLES ............................................................................................................ 9 LIST OF FIGURES ........................................................................................................ 10 ABSTRACT ................................................................................................................... 11 CHAPTER 1 INTRODUCTION .................................................................................................... 13 Ecological and Human Dimensions of Conflicts between Birds and Crop Production ........................................................................................................... 13 Conflicts between Monk Parakeets and Crop Production ....................................... 15 Research Overview ................................................................................................ 17 2 MONK PARAKEET ABUNDANCE AND DAMAGE TO CROP FIELDS IN RELATION TO LOCAL AND LANDSCAPE VARIABLES ....................................... 18 Background ............................................................................................................. 18 Methods .................................................................................................................. 21 Study Area ........................................................................................................ 21 Sampling Scheme ............................................................................................ 22 Bird Abundance Surveys .................................................................................. 23 Estimation of Crop Damage ............................................................................. 24 Within field and Field level Variables ............................................................... 26 Landscapelevel Variables ............................................................................... 27 Statistical Analyses .......................................................................................... 30 Results .................................................................................................................... 31 Monk Parakeet Abundance and Damage in Crop Fields .................................. 31 Monk Parakeet Abundance and Damage in Crop Fields in Relation to Within field Field and Landscape Variables ................................................. 32 Discussion .............................................................................................................. 34 Factors Related with Monk Parakeet Abundance and Damage to Crop Fields ............................................................................................................ 34 Management Implications ................................................................................. 39 3 DENSITY OF MONK PARAKEET NESTS IN SITES WITH EUCALYPTUS IN RELATION TO LOCAL AND LANDSCAPE VARIABLES ....................................... 48 Background ............................................................................................................. 48 Methods .................................................................................................................. 51 Study Area ........................................................................................................ 51

PAGE 7

7 Sampling Scheme ............................................................................................ 52 Nest Abundance Surveys ................................................................................. 53 Local Level Variables ....................................................................................... 54 LandscapeScale Variables .............................................................................. 55 Statistical Analyses .......................................................................................... 57 Results .................................................................................................................... 58 Density of Parakeet Nests and Characteristics of Nest Patches and Landscapes ................................................................................................... 58 Density of Monk Parakeet Nests in Relation to Local and Landscape V ariables ....................................................................................................... 59 Discussion .............................................................................................................. 60 Factors Related with Density of Monk Parakeet Nests ..................................... 60 Management Implications ................................................................................. 64 4 INFLUENCE OF SOCIO PSYCHOLOGICAL AND SOCIO DEMOGRAPHIC FACTORS ON FARMERS PREFERENCES FOR MANAGEMENT STRATEGIES TO DECREASE MONK PARAKEET DAMAGE T O CROPS ........... 70 Background ............................................................................................................. 7 0 Methods .................................................................................................................. 74 Study Area ........................................................................................................ 74 Sampling Scheme and Questionnaire Design .................................................. 75 Variable Measurement ..................................................................................... 76 Preference for management strategies ...................................................... 76 Socio psychological factors ........................................................................ 77 Socio demographic factors ........................................................................ 81 Statistical Analyses .......................................................................................... 82 Preference scores and reliability ................................................................ 82 Assessment of the relationship between pr eferences for management strategies and sociopsychological and socio demographic factors ....... 83 Results .................................................................................................................... 84 Preferences for Bird Pes t Management Strategies .......................................... 84 Factors Related to Preferences for Management Strategies ............................ 85 Socio psychological factors ........................................................................ 85 Socio demographic factors ........................................................................ 90 Relationships among Sociopsychological and Sociodemographic Factors .... 91 Results Summary ............................................................................................. 92 Discussion .............................................................................................................. 93 Preferences of Farmers for Management Strategies and Factors Related with those Preferences .................................................................................. 93 Management Implications ................................................................................. 97 5 CONCLUSIONS ................................................................................................... 108 Influence of Local and Landscape Variables on Monk Parakeet Abundance or Damage in Crop Fields and Nesting Sites ......................................................... 108

PAGE 8

8 Human Dimensions of Conflicts between Monk Parakeets and Crop Production 111 Management Implications ..................................................................................... 113 APPENDIX A COMPLEMENTARY RESULTS FROM CHAPTERS 2 AND 3 ............................. 117 B QUESTIONNAIRE STRUCTURE AND CONTENT .............................................. 122 C QUESTIONNAIRE IN SPANISH ........................................................................... 124 D MOSAIC GRAPHICS FO R RELATIONSHIPS AMONG INDEPENDENT VARIABLES .......................................................................................................... 138 LIST OF REFERENCES ............................................................................................. 141 BIOGRAPHICAL SKETCH .......................................................................................... 156

PAGE 9

9 LIST OF TABLES Table page 2 1 Minimum AICc models and regression results for factors used in predicting monk parakeet abundance in crop fields in Paran department ......................... 45 2 2 Minimum AICc models for monk parakeet damage to corn and sunflower fields in Entre Rios (Argentina) during 2007 and 2008 summer seasons. .......... 46 2 3 Regression results for factors considered in predicting monk parakeet damage to crop fields in Paran department (Entre Ros, Argentina) in 2007 and 2008 summer seasons. ............................................................................... 47 3 1 Summary o f vegetation structure and landscape metrics included in models of monk parakeet nest density in inhabited farms with eucalyptus in east central Argentina. ............................................................................................... 67 3 2 Minimum AICc models for densi ty of monk parakeet nests in inhabited farms with eucalyptus trees in east central Argentina (n=35). ...................................... 68 3 3 Regression results for factors considered for predicting density of monk parakeet nests in inhabited farms with eucalyptus trees in east central Argentina. ........................................................................................................... 69 4 1 Summary of regression results for sociopsychological and sociodemographic variables and preferences for managem ent strategies. .............. 105 4 2 Top performing variables related to preferences of farmers for management strategies to decrease monk parakeet damage to crops based on the AIC value and percent concor dance, which represents the association of predicted probabilities and observed responses .............................................. 106 4 3 Results from bivariate Chi square test evaluating the correlation between independent variabl es. ..................................................................................... 107 A 1 Correlations between landscape metrics used to quantify landscape composition and configuration around focal fields. ........................................... 118 A 2 Suite of models used to describe the relative abundance and damage of monk parakeet to crop fields at withinfield, field and landscape levels in Paran (Entre Ros, Argentina), 20072008. .................................................... 119 A 3 Correlations between landscape metrics used to quantify landscape composition and configuration around nesting patches. ................................... 121

PAGE 10

10 LIST OF FIGURES Figure page 2 1 Map showing the location of Department of Paran (Entre Ros Province, Argentina) and the crop fields sampled in 2007 and 2008. ................................. 41 2 2 Sampling scheme for evaluating bird damage to crop fields. ............................. 42 2 3 Monk parakeet damage to corn ears. ................................................................. 43 2 4 Monk parakeet damage to sunflower heads. ...................................................... 43 2 5 Buffer extents used for sampling landscape level variables from satellite images around each crop field (central point). .................................................... 44 3 1 Buffer e xtents used for sampling landscape level variables from satellite images around each nesting patch (central point). ............................................. 66 4 1 Ranking of preference for management alternatives ( s.e.) by farmers .......... 101 4 2 Distribution of the levels of preferences for each management strategy based on a multiple correspondence analysis (MCA). ................................................ 102 4 3 Frequency of farmers knowing about management alternatives for decreasing monk parakeet damage before the interviews. .............................. 103 4 4 Frequency of farmers reporting management alternatives for decreasing monk parakeet damage as the most effective one among all known alternatives. ...................................................................................................... 104 D 1 Mosaic display showing the relationships among age, education and attitudes toward monk parakeets. .................................................................... 138 D 2 Relationships among attitudes and beliefs about effectiveness of management strategies. ................................................................................... 139 D 3 Relationships between beliefs about the most effective management strategy and influence of subjective norms. Management strategy: ............................... 140

PAGE 11

11 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 ECOLOGICAL AND HUMAN DIMENSIONS OF THE MONK PARAKEET DAMAGE TO CROPS IN ARGENTINA By Sonia Beatriz Canavelli December 2011 Chair: Ly n Branch Major: Wildlife Ecology and Conservation The Monk Parakeet ( Myiopsitta monachus ) is considered among the most important bird pest species causing damage to crops in Argentina. In this study, I explored habitat features influenc ing abundance and damage of monk parakeets to crop fields and density of nests in inhabited farms with eucalyptus trees at multiple spatial levels Additionally I examined socio psychological and socio demographic factor s influenc ing farmers preferences about management of monk parakeet damage to crops. Monk parakeet abundance and damage was greater in sunflower than in corn fields. Landscape variables such as dista nce to nearest sites with trees, percentage of landscape with trees and availability of fo raging sites for m onk parakeets around the crop fields were more important than local variables in explaining monk parakeet damage to crop fields. However, local variables, such as field area, p lant density and percentage of field border with trees also were related to d a mage C onversely, the density of monk parakeet nests in inhabited farms with eucalyptus trees was not clearly explained by any variable or combination of variables modeled in this study

PAGE 12

12 Farmers preferred population control strategies, such as nest destruction and killing of birds, for decreasing monk parakeet damage to crops Preferences of farmers for management strategies were related more strongly to attitudes toward monk parakeets than to any other factor considered in this study Other important soci o psychological factors were perceived efficacy and p revious knowledge about management strategies P erceptions of magnitude of damage by monk parakeets practically were not related to preferences S ocio demographic factors such as age and education, were related to preferences in different ways depending on the management strategy Based on this study, m anagers should consider both local and landscape factors when planning management measures to prevent monk parakeet damage to crop and reduce nesting on f arms. Additionally, extension actions should be oriented to modifying attitudes toward monk parakeets as well as communicating and showing the efficacy of alternative management strategies. Given the current uncertainties in the outcome of management actio ns, an adaptive management approach would be useful to evaluate the efficacy of strategies other than lethal or reproductive control

PAGE 13

13 CHAPTER 1 INTRODUCTION Ecological and H uman D imensions of C onflicts between B irds and Crop P roduction Conflicts between wildlife and human activities have existed historically. However, the number and severity of conflicts have increased and are expected to continue to increase in the future as a consequence of human population growth (Fall and Jackson 2000; Messmer 2000; Bruggers et al 2002; Linz et al. 2002). Additionally, public attitudes towards of wildlife are changing, with increasing public concern over the welfare of animals, including overabundant species ( Messmer 2000; Bruggers et al. 2002). As a consequence, the need for effective, environmentally safe and sciencebased management methods and strategies probably is more critical today than in the past. Underlying t he development of efficient management strategies to prevent and/or decrease wildlife damage is a go od understanding of species behavior and ecology and the behavior of people in response to the damage and its management (C onover 2002). G ranivorous bird species associated with agroecosystems cause damage to crops, feedlots and stored grains worldwide (P inowski and Kendeigh 1977; De Grazio 1978; Feare 1993 ; Bruggers and Zaccagnini 1994). The number of species causing this damage is relatively small, but their impacts often are significant ( Pinowski and Kendeigh 1977; De Grazio 1978; Feare 1993 ; Bruggers e t al. 1998). Some species, such as red winged blackbirds ( Agelaius phoeniceous ) in North America, eared doves ( Zenaida auriculata) in South America and red billed quelea ( Quelea quelea) in Africa, comprise flocks and communal roosts of many thousands of individuals and range widely (Beletsky 1996; Bucher 1992a; and Bruggers and Elliot 1989, respectively). Other species that cause damage, including the roseringed ( Psittacula krameri ) and the

PAGE 14

14 monk parakeets ( Myopsitta monachus ) a lthough less numerous and more sedentary are also very social and visible (Spreyer and Bucher 1998; Kahn 2003). The conspicuousness of the se birds and the damage they cause, plus the high variability in damage, make objective estimation of damage by farmers difficult (Conover 2002) and contribute to a tendency to overestimate losses (Dyer and Ward 1977; Bucher 1992a ; 1998). A s a result, farmers often apply management measures to decrease bird damage that are not economically effective or are contrary to research findings (Bomford and Sinclair 2002; Tracey et al 2007). The deficiencies observed when planning management strategies to decrease bird damage to crops, as well as the lack of effectiveness of many of t hese strategies have been attributed to failures to consider the human dimensions of the problem including sociological and psychological aspects ( Timm 1991 in Clergeau 1995; Bomford and Sinclair 2002). T he biology, physiology and ecology of some vertebrate pests species, including birds, is relatively wellknown and, for thi s reason, the failures of management programs often are due to the lack of considerations of human dimensions (Timm 1991 in Clergeau 1995) However, the dynamics of species distribution are not always well known, particularly at landscape scales, when related to problems with birds in crop fields (Clergeau 1995) Birds move at greater scales than individual properties of particular crop fields, using both cultivated and noncultivated fields in their life cycles (e.g., starlings, Bruun and Smith 2003; blac kbirds, Beletsky 1996 ; Orians 1985). The availability of cultivated site and alternative foraging locations around a crop field, as well as uncultivated sites used for roosting and/or breeding by bird pest species influence bird abundance and

PAGE 15

15 damage on particular crop fields (Otis and Kilburn 1987; Tourenq et al. 2001; Amano et al. 2004, 2008; Hagy et al. 2008). Additionally, the characteristics of the crop field, such field area, plant density or weed abundance, also could influence the abundance and damage by birds on particular crop fields (Otis and Kilburn 1987; Tourenq et al. 2001; Amano et al. 2004, 2008; Hagy et al. 2008). Therefore, studying the ecology of bird pest species r equires the consideration of multiple scales of observation, from particular fields (or patches) to the surrounding landscape or region (Clergeau 1995). However, with exception of few works ( Buch er 1990; Bucher and Ranvau 2006, Cavallero 2010) practically no studies had considered multiple scales of observation with neotropical bird pest species in South America. Conflicts between M onk P arakeets and C rop P roduction The monk parakeet is a medium size ( 90120 g ) neotropical parrot species commonly involved in humanwildlife c onflicts in its native range (South America) and nonna tive areas of distribution (North America and Europe, Spreyer and Bucher 1998). In Argentina and Uruguay, the monk parakeet is among the most important bird pest species causing damage to grain crops (Bucher and Bedano 1976; Bucher 1984, 1992a, b; Bruggers y Zaccagnini 1994; Bruggers et al. 1998). This species also cause s damage i n other settings (e.g. fruit crops and electric utility structures, Bucher 1992a; Bucher and Martin 1987). In North America and Europe, the monk parakeet primarily causes problems i n urban settings related to location of nests, including damage to electric utility structures that lead to power outages ( Avery et al. 2002) and disturbance to tranquility of human neighborhoods, due to the noise parakeets produce (Santos 2005; Burger a nd Gochfeld 2009). Currently the monk parakeet is a threat to agriculture production or native biota in some nonnative areas but if its populations

PAGE 16

16 continue to expand, this species could cause more problems in the future (Sol et al. 1997; Tillman et al. 2000; Domenech et al. 2003). Traditionally, lethal control has been preferred as the most effective method for decreasing monk parakeet crop conflicts particularly in Argentina (Bucher 1984, 1992). Several methods have been used, including nest burning or destruction, shooting, payment of bounties, trapping, netting, toxic baits, and spraying of nests with insecticides. Since the 1980s, the primary lethal control method has been insecticides mixed with grease and applied on nest openings to produce intoxi cation and potentially death of birds entering to the nest (Arambur 1991). However, objections to this method are increasing and new methods are required (Canavelli and Zaccagnini 2007 ; Canavelli and Arambur in press ). Additionally, monk parakeets represent conflicting values for different groups of people because, although they are considered a pest species, this species also is valued as a domestic pet (Moschione and Banchs 2006). Currently, most information about monk parakeets biology and ecology in Argentina is focused in population demographics and social behavior (Bucher et al. 199 1 ; Arambur 1991 ; Navarro et al. 1992 ; Eberhard 1998). However, the available information is relatively scarce, particularly in aspects such as habitat use (Spreyer and Bucher 1998). Consequently, understanding habitat features influencing the use of crop fields and nesting sites for by monk parakeets in agricultural landscapes of Argentina may help the management of conflicts with this species, not only in Argentina but i n other native and n onnative areas of distribution. In addition, research o n the human dimensions of the problem could improv e the comprehension of the social situation in which management occurs and, consequently, contribute to increased success of

PAGE 17

17 ma nagement strategies in the future. Currently, no information is available about social and psychological factors underlying farmers preferences for management strategies to decrease conflicts with monk parakeet s (Canavelli and Zaccagnini 2007). Research O verview The goal s of this study w ere to: 1) identify habitat features influencing the abundance of monk parakeets in crop fields and damage by parakeets to crops 2) identify habitat features that influence abundance of nests and 3) e xamine factors in fl uenc ing the decisionmaking process by farmers about management of monk parakeet damage to crops. Because habitat selection involves a hierarchy of decisions at multiple scales (Hildn 1965 ; Johnson 1980 ; Cody 1985; Wiens et al. 1987), I used a multilevel approach for analyzing local and landscape factors influencing the abundance and damage of monk parakeets in corn and sunflower fields and the density of monk parakeet nests in sites with e ucalyptus trees in central Argentina. Chapter 2 focuse s on monk parakee t damage to crop fields, while C hapter 3 focuses on density of monk parakeet nests in sites with e ucalyptus trees. Chapter 4 focuses on the human dimensions of conflicts with monk parakeets and crop production. I applied a behavioral decision approac h to determine farmers preferences for management strategies to decrease damage from monk parakeets to crops in Argentina and to evaluate sociopsychological factors and sociodemographic factors that influenced those preferences Fin ally, in C hapter 5, I summarize the results and management implications from the previous chapters. This study is one of the first studies applying a multi level analysis to understand monk parakeet ecology in agricultural landscapes and to my knowledge; it is the first study to apply a human behavioral model in bird pest management worldwide.

PAGE 18

18 CHAPTER 2 MONK PARAKEET ABUNDANCE AND DAMAGE TO CROP FIELDS IN RELATION TO LOCAL AND LANDSCAPE VARIABLES Background Bird damage to agricultural crops is a cause of economic loss for far mers in many parts of the world (De Grazio 1978; Conover 2002). Understanding the underlying factors favoring damage on specific sites is crucial for predicting occurrence of damage and, consequently, focusing resources to reduce damage on sites and/or tim es where damage is more likely (Amano et al. 2008). Additionally, understanding factors favoring use and damage of a particular crop field by birds could help with design and evaluation of science based management strategies for preventing damage, such as alternative feeding areas or lure crops (Amano et al. 2007 ; Hagy et al. 2008). Factors influencing use and damage of a particular crop field by birds act at multiple scales ( Clergeau 1995). For instance, local characteristics related to the crop field, such as crop structure (e.g., plant density and height) and weed density have been found to be related to damage by redwinged blackbirds ( Agelaius phoeniceus ) in sunflower fields (Otis and Kilburn 1987). However, characteristics of the landscape around the sunflower fields, such as availability of nesting or alternative food habitats near the crop field, may be more important than local characteristics for explaining blackbird damage to sunflower fields (Otis and Kilburn 1987; Hagy et al. 2008). Landscape ch aracteristics also have been found to be important predictors of damage to crops for other bird pest species, such as flamingos ( Phoenicopterus rubber roseus ) in rice fields (Tourenq et al. 2001) and whitefronted geese ( Anser albi frons ) in wheat fields (A mano et al. 2004, 2008).

PAGE 19

19 Agricultural landscapes are a mixture of cultivated and uncultivated patches (fields), varying in composition (i.e., amount of land cover type s in the landscape) and configuration (i.e. spatial arrangement of patches within the l andscape) at multiple spatial and temporal scales (Forman and Godron 1986; Burel and Baudry 1995; Holt et al. 1995; Landis and Marino 1999). Mobile species using these landscapes, including b ird species causing damage to crops often use cultivated and non cultivated patches in their life cycles. The abundance and distribution of these cultivated and noncultivated patches in the agricultural landscape may influence the abundance and damage of a bird species o n a particular patch or crop field ( Otis and Kilburn 1987 ; Tourenq et al. 2001; Amano et al. 2004, 2008; Hagy et al. 2008) Multi level studies in which researchers evaluate the influence of local and landscape variables on animal abundance and distribution in particular patches ( or focal patches, B rennan et al. 2002) have been proposed as a helpful tool for integrating multiple scales of observation about the behavior of bird pest s in agricultural landscapes and designing rational management strategies (Clergeau 1995). Such multi level studies are c ommon in landscape ecology and conservation biology ( see review s in Mazerolle and Villard 1999 and Thorton et al. 2010) However, few studies evaluating bird pest damage to crop fields have explicitly addressed multiple scales of observation in the same st udy (but see Otis and Kilburn 1987; Tourenq et al. 2001; Amano et al 2004, 2008 ; Hagy et al. 2008). In this study, I applied a multi level approach to analyze factors influencing the abundance and damage of monk parakeet s ( Myiopsitta monachus ) to crop fields. The monk parakeet is among the most important bird pest species causing damage to grain

PAGE 20

20 crops i n South America particularly in Argentina and Uruguay (De Grazio and Besser 1975; Bucher and Bedano 1976; Bucher 1992 a,b ; Bruggers et al. 1998). Damage to grain crops by monk parakeets occurs principally to ripening sunflower and corn and occasional ly sorghum, wheat and rice (Spreyer and Bucher 1998) Quantification of damage to crops by monk parakeets is very scarce, but indicates moderate (< 5%) to high (up to 20%) crop loss (De Grazio 1985; Bucher 1992a; Canavelli et al. 2008). Key habitat elements for parrots, including monk parakeets are suitable nesting sites and highquality food (del Hoyo et al. 1992). In agricultural landscapes, suitable nesting si tes for monk parakeets are varied because, unlike other parrots, the monk parakeet does not nest in cavities but rather constructs nests with sticks on tall natural and artificial structures, including native savanna trees (e.g., Prosopis spp. and Acacia spp.), introduced Eucalyptus trees, and utility poles (Spreyer and Bucher 1998). H igh quality foods for monk parakeets in agricultural landscapes ar e maturing grain crops that they prefer, such as sunflower and cor n (Spreyer and Bucher 1998; Ara mbur 199 7, 199 8 ; Arambur and Bucher 1999). However, parakeets also forage on wild seeds, fruit of native trees, and other grain and fruit crops (Spreyer and Bucher 1998) M onk parakeets are central place foragers ( Stephens and Krebs 1985) because they use the nest all year around, both for breeding and roosting. They forage out from the nest and then return to that site D aily movement f rom the nest site to foraging areas is generally between 3 and 5 km, although possibly longer (up to 24 km) during the non breeding season ( Spreyer and Bucher 1998). Considering the wide range of daily movement of monk parakeets I expected characteristics of the landscape around a particular crop field would influence the

PAGE 21

21 abundance and damage of parakeets in that particular field. S pecifically, given that m onk parakeets have a generalist foraging behavior (Bucher et al. 1991; Hyman and Pruett Jones 1995) I expected availability of alternative foods on the landscape around a crop field to be related to abundance and damage of monk parakeets in the field. Additionally, because monk parakeets use trees for perching, nesting, or daily loafing I expect ed abundance and distribution of patches with trees around a given crop field to be important for explaining the abundance and damage by m onk parakeet s i n the field. Methods Study A rea The study was conducted in a 525,000ha area comprising the Department of Paran (Entre Ros Province, Argentina, Figure 2 1). The area is characterized by diverse production activities, with a predominance o f crops, beef cattle and milk production (Engler and Vicente 2009). Agricultural crops cover about 49% of the area, including in order of importance soybeans, wheat, corn, sunflower and sorghum. The Department of Paran contains approximately 15% of the total agricultural area in the province and is the most important department in Entre Ros in this respect (Engler and Vicente 2009). Mean annual temperature is 19C (12C in winter and 25in summer) and mean rainfall is approximately 1000 mm. A gradient in production activities and, therefore landscape pattern, occurs in the study area from northeast to southwest. The northeast mostly is devoted to a mixture of crops and cattle production, with high interspersion of woodlands, pastures and crops. The sou th west is intensively agricultural, with a few patches of woodland interspersed with large fields of annual crops.

PAGE 22

22 Sampling S cheme The study was conducted in the 2006200 7 and 2007200 8 austral summer seasons (December to February) Most damage to crops by monk parakeets occurs in summer, following their spring reproductive period (August December, Bucher 1992a). I used a Geographic Information System (ArcGIS v.9.2) to place a 10x10km grid over Paran Department and selected 25 noncontiguous grid cells in 200 6 and 31 cells in 200 7 using systematic sampling with random start (first cell selected at random and every other cell selected thereafter) Taking the geographic coordinates for the central point of each cell as a reference in the field, I identifi ed the most proximate corn or sunflower field to that point. Based on the type of crop I sampled on the first cell, when possible I choose a different type on the next cell in order to have both types of crop fields with a relatively even distribution throughout the study area ( Figure 2 1) A crop field (or patch) was defined as a contiguous area covered by corn or sunflower, differing from its surroundings Based on a first visit to each crop field, I planned the date for sampling bird abundance and damag e to coincide with the ripening crop in each field, which is when damage by monk parakeets was expected. Study sites included 14 corn and 11 sunflower fields in the 20062007 summer season (hereafter 2007 season) and 15 corn and 26 sunflower fields in the 2007200 8 summe r season (hereafter 2008 season). Because of problems during field sampling (early harvest, immature stage) one corn field sampled in 2007 and one sunflower field sampled in 2008 were eliminated from the data pool Additionally, d ata from t wo fields (1 corn and 1 sunflower field) that were consistently outside the distribution range of values for all independent variables were eliminated from the pool of data, as well as three fields (1 corn and 2

PAGE 23

23 sunflower fields) with missing values for at least one variable. Therefore, final sample size for statistical analyzes was 22 corn fields and 27 sunflower fields. The mean size ( s.e.) of corn fields was 22.52 ha ( 3.46) and the mean size of sunflower fields was 24.25 ( 2.99 ). Bird A bundance S ur veys Monk parakeets were surveyed using unlimiteddistance 5min point counts (Bibby et al. 2000; Freemark and Rogers 1995). Because monk parakeets were difficult to observe once they entered the crop field and estimates of distance were difficult on the h omogenous surface of a field, I used point counts with 180 semicircles of unlimited distance in direction o f the crop (Bibby et al. 200 0 ; Freemark and Rogers 1995). I used number of birds observed/point/plot as a n index of bird abundance in the field Al l parakeets observed in the field as well as entering or leaving the plot were recorded. The number of parakeets was counted for individual birds or small groups, or estimated otherwise. Points were located on the border of the crop fields in proportion to their size, considering a minimum distance of 200 m between consecutive points to decrease the possibil ity of double counting birds (Freemark and Rogers 1995; Boutin et al. 1999a and 1999b; Best et al. 2001). Surveys were conducted between sunrise and m id morning ( 10:00 h), with one field sampled per morning. The same observer conducted all point counts in 2007 In 2008, another observer with experience i n bird counts in crop fields was included and observers were randomly assigned to crop fields. Relative abundance of monk parakeets was estimated for each crop field as t he average number of birds observed per point per field.

PAGE 24

24 E stimation of C rop D amage E ach crop field was sampled for monk parakeet damage in a fixed number of small plots (n= 36 in 2007 and n=80 in 2008) distributed along transects in the field ( Figure 2 2, Otis 1992 ; Zaccagnini 1998 ; Linz 1999) Based on the size and shape of the field, the field was divided i n 2 4 s ections containing an equal number of row s. A row was randomly selected i n the first section and the rows for the other s ection s were placed at a fixed distance from each other so that sampled rows were systematically distributed over the width of the field. T hree strata were sampled perpendicular to each sample row : field edge (first line with crop plants), border (25 m from the edge of the field) and center of field Sample plots on the field edge corresponded to plants on the first line, and sample plots on the border and the center sections of each row were systematically pla ced with a random start in order to have a fixed number of samples per stratum per field (8 plots in field edge, 16 plots in border and 12 plots in center in 2007 and 12 plots in field edge, 36 in border and 32 in center in 2008, Figure 2 2). In 2008, four of the 12 plots in the field edge and twelve of the 36 plots in the border were taken in four additional short transects (25 m) in one border of the field in an attempt to increase accuracy in damage estimations. Each plot consisted of 5plants perpendic ular to the direction of the sampling row ( Figure 2 2). In 2007, I registered the number of damaged and nondamaged plants for each plot (infestation or frequency of damage). Additionally, in 2007 I visually estimated intensity of damage (i.e., percentage of grain loss) on damaged plants as the damaged length of corn ear (De Grazio et al. 1969) and the percentage of sunflower head damaged (Dolbeer 1975; Zaccagnini and Cassani 1985; Zaccagnini and Tate 1992; Otis 1992). Sunflower plants always had one head p er plant. T he cases where I ha d 2

PAGE 25

25 damaged corn ears in a plant (n=9 cases), damage was averaged between both corn ears in the plant. In corn fields, damage of ears was attributable to monk parakeets when nosigns of mammal activities, such as tracks or feces, were observed in the field, the external cover of the ear was opened and the ear physically damaged (for example, with the top part of the ear, Figure 2 3). Monk parakeet is the only bird species capable of producing this type of damage in the region (Bucher and Bedano 1976; Bucher 1992a). In sunflower fields, I differentiated damage by monk parakeets from other birds ( doves and pigeons) based on the existence of physical damage on the head (e.g., some parts missing) in addition to husks from sunflower seeds on the plant or in the ground ( Figure 2 4). B ecause of the small magnitude of damage intensity and the high variation among fields ( =0.18% grain loss, SD =0.32 in corn and =0.91 % of grain loss SD =1.96 in sunflower), as well as a direct relationship between the frequency and intensity of damage (R2=0.67, p= 0.09 for corn, R2=0.76, p<0.001 for sunflower, see Canavelli et al. 2008 for more details), I only evaluated frequency of dam age in 2008, increasing effort in each field from 36 to 80 plots per field i n an attempt to increase precision i n the estimator Frequency of d amage by monk parakeet s i n each field was estimated as the propor tion of damaged plants over the total number of plants using a stratum weighted proportional estimator (Cochran 1977 ; Zaccagnini et al. 1983, 1985). The number of plants by stratum was estimated based on plant density (number of plants per square meter, es timated as the number of plants per meter of row divided by row distance in meters and multiplied by a square meter, O. Valentinuz, pers.com.) and

PAGE 26

26 the surface of each stratum in square meters was calculated using Patch Analyst extension in ArcGIS 9.2 (Remp el 2010). Within field and Field level V ariables V ariables that characterized crop structure within each crop field were chosen based on previous studies indicating the influence of these variables on the use of crop fields by bird pests (Otis and Kilbur n 1988; Hagy et al. 2008). S ampling plot s used for damage evaluation were used for measuring these within field variables (n= 36 in 2007 and n=80 in 2008) In each plot (5 plants each) I recorded plant height and plant phenological stage for 3 of the 5 pl ants in the plot (one in the center and one at each extreme of the plot). Additionally, I measured number of rooted plants/ row meter and row width (both variables related to plant density) from the plant in the center of the plot and visually estimated weed coverage as the proportional coverage of a 1x1m quadrat in each sampling plot (Otis and Kilburn 1988; Colbach et al. 2000). Measurements for crop structure variables at each plot were then averaged over all sampling plots in a field to obtain one value per field for each variable. Because plant density and plant h plants, I used plant density for model construction (Statistical analyses). Field level variables characterized the field as a patch within the landscape. The percentage of the field border with trees was recorded in the field on a 3point scale (1= 0 5%, 2= 550% or 3=>50%) as well as the presence of crops, pastures, weedy fields or woodland adjacent to the field. Based on field observations and discussions with each landow ner, I determined that no control measures were taken against monk parakeets on the crop fields evaluated in this study. All crop fields (focal fields) were digitized using Google Earth and the geographic coordinates of crop borders recorded in the

PAGE 27

27 field. Polygons were then converted to a vector file and imported in ArcGIS (v 9.2). Using Patch Analyst extension for ArcGIS (Rempel 2010), I estimated field area, perimeter, and shape complexity (as a shape index). Because field area and perimeter were substan (Statistical analyses) based on the relationship between bird damage and field area, either negative (e.g., for redwinged blackbirds, Clark et al. 1982; Zaccagnini and Dabin 1985) or pos itive (e.g., for greater flamingos Tourenq et al. 2001) as well as its simplicity for use and evaluation in the field. Landscapelevel V ariables I evaluated landscape context around each crop field using both distancebased measures and buffer measures of landscape composition and configuration. Using Google Earth, I measured distance from the crop field to the nearest site with manmade structures, such as houses and barns, and trees, which are commonly used by monk parakeets as nesting sites (Chapter 3) Additionally, I examined the composition and configuration of the landscape within circular buffers of 3 different radii from the center of each crop field (1000, 3000 and 5000 m Figure 2 5 ) These landscape extents were chosen based on the expected daily movement range of monk parakeets from the nest site to foraging areas while breeding (range: 3.58 km, mode: 35 km, Spreyer and Bucher 1998). I set an upper buffer limit of 5000 m to avoid the problem of overlapping buffers and potential spatial aut ocorrelation of the local landscapes around each crop field (Koper and Schmiegelow 2006; Renfrew and Ribic 2008; Boscolo and Metzger 2009). I did not use buffers smaller than 1000 m because of problems with artificial borders i n estimation of landscape indices.

PAGE 28

28 Buffers for crop fields sampled in 2007 and 2008 were obtained from Landsat TM images (22682 21Jan 2007 and 24Jan 2008) classified by Noelia Calamari (INTA, EEA Paran). Both images were classified using supervised classification. The 2007 Landsat image was classified using ECHO (Extraction and Classification of Homogeneous Objects) in MultiSpect Application v3.1 (2007) and the 2008 Landsat image was classified using ImageSVM (Support Vector Machine, van der Linden et al. 2009) in ERDAS imagine 9.1 (2006). Ten land cover types were identified: water; corn; sunflower; soybeans ; sorghum; pastures and other agricultural uses (e.g., fallow and weedy fields); developed areas, plowed and some fallow fields (which could not be clearly distinguished); introduced trees; native trees; and riparian vegetation. Results from each classification were validated with 100 points per land cover type randomly selected using Quickbird images (available in GoogleEarthTM, http://earth .google.com ) and ground sampling. Overall classification accuracy was 82 % and 84 % for 2007 and 2008 satellite images, respectively. Finally, to clearly distinguish among different land cover types and decrease the problem of artificial borders with rast er images, I re grouped the land cover types into 7 classes: water; crops susceptible to damage by monk parakeets (corn and sunflower); nonsusceptible crops (soybean and sorghum, immature at the time the image was obtained); pastures and other agricultural uses (e.g. fallow and weedy fields); developed areas, plowed and some fallow fields (which could not be clearly distinguished); and tree patches, including native and introduced trees. I focused the analysis on the availability of three land cover classes: 1) crops that could be susceptible to damage by monk parakeets (corn and sunflower), 2) tree patches, potentially used as primary sites for perching, nesting or daily loafing, and 3)

PAGE 29

29 pastures and other agricultural uses (e.g., fallow and weedy fields), which can include food items for monk parakeets such as flowers and seeds I used FRAGSTATS 3.3 software (McGarigal et al. 2002) to calculate landscape metrics representing landscape composition and configuration. Composition metrics included percentage of landscape with corn and sunflower, other agricultural uses (including pastures, fallow and weedy fields) and native and introduced trees. Configuration metrics included the aggregation of susceptible crops and trees within each buffer ( measured with a cl umpiness index ) m ean nearest neighbor distance among attractive crops and tree patches considering all attractive crops or tree patches on the landscape, respectively, and patch shape complexity (measured with a shape index) of attractive crops and tree patches. Substantial correlations (r among some landscape metrics, particularly at higher extents (3000 and 5000 m, Table A 1, Appendix A ) and only uncorrelated metrics ( r<0.60) were included in the same model. Because percentage of landsc ape with different land cover types could be important for explaining bird abundance on a site ( Fahrig 2001; Renfrew and Ribic 2008; Hagy et al. 2008), for analyses I sought to include configuration metrics uncorrelated with percentage of landscape for the two primary cover classes (crops susceptible to damage and trees). In the case of susceptible crops, this was possible with the clumpiness index. However, in the case of tree patches, all configuration metrics were correlated with percentage of this cover type on the landscape. Percentage of the landscape with tree patches was correlated negatively with mean nearest neighbor distance and positively with patch shape complexity and clumpliness Therefore, I only considered percentage of the landscape with tr ees for data analyses.

PAGE 30

30 Finally, because percentage of landscape with crops susceptible to damage was negatively correlated with percentage of landscape with trees at higher buffer extents (3000 and 5000 m, Table A 1, Appendix A), I included only one of them at a time for model construction at those buffer extents. Statistical A nalyses I modeled relative abundance of monk parakeets and monk parakeet damage in each crop field as a function of withinfield, field and landscape variables at each buffer extent (1000, 3000 and 5000 m) separately. Additionally, I constructed a set of models combining withinfield, field and landscape variables. The response variables (relative abundance of monk parakeets and damage) were examined for deviation from normality (Infostat, Di Rienzo et al. 2010, SAS v. 8.0; SAS Institute Inc., 2006). Because both variables were no t normally distributed, nor were transformations of those variables (e.g., square root for abundance or cosine for proportion of damage) normally distributed, a generalized linear model framework (GLM) was selected for modeling purposes, with a negative binomial error structure for relative abundance of monk parakeet and a binomial error structure for proportion of crop damage. Because final sample sizes for model construction were relatively small for each crop (n=22 for corn and n= 27 for sunflower), each model was restricted to include between one and three variables. I ran models for all single variables, all sets of two variables and then all sets of 3 vari ables within each level (withinfield, field and landscape, Table A 2 Appendix A). Then, for parakeet damage, I ran multi level models that contained the strongest predictors from each of the three levels. I used the within field variable with the minimum AICc value as the base model for adding the best performance field and

PAGE 31

31 landscape variables in the multi level models (Fletcher and Koford 2002; Renfrew and Ribic 2008). Multi level models were not run for parakeet abundance because I was not able to clear ly identify predictors within each level. I developed models for corn and sunflower separately, to be confident that responses at field or landscape levels were not driven by response to field quality (i.e., crop type). These crops differ in withinfield variables (e.g., plant height, plant density, weed density, Oscar Valentinuz, pers.com.). Additionally, I expected the use, and potentially damage, of crop fields to differ based on differential preferences for these crops ( Arambur y Bucher 1999). All mod els were evaluated using SAS PROC GENMOD and maximum likelihood estimation. I used Akaike Information Criteria adjusted for small sample size (AICc) for comparing model performance, and I restricted AICc scores to 2 for model retention (Burham and Anderson 2002). For evaluating individual variable performance at each level, I used model averaging and the sum of i, Burham & Anderson 2002). Result s Monk P arakeet A bundance and D amage in C rop F ields Monk parakeets differentially used crop fields based on the crop type. Monk parakeets were significantly more abundant in sunflower than in corn fields ( x = 9 29 monk parakeet/point/field, SE = 1 5 2 and x = 5 27 SE=1.6 1, respectively, Wilcoxon test = 549.50, p=0.007), and no statistically significant differences occurred in abundance within each crop type between years ( Wilcoxon test= 171.00, p=0.11 for corn; Wilcoxon test= 205.50, p= 0 .22 for sunflower ). Monk parakeets were observed in 29 of the 30 fields with sunflower (038 parakeets observed per point per field), and in 14 of the 24 corn fields ( 0 33 monk parakeets observed per point per field).

PAGE 32

32 Similar to abundance, monk parakeet damage differed according to the crop type. Damage was significantly higher in sunflower than in corn fields ( x = 4 29 % damaged plants SE= 0 88; x = 0 90, SE= 0 4 6 respectively, Wilcoxon test= 455.00, p <0.0001), and no statistically significant differences occurred in damage to each crop type between years ( Wilcoxon test= 166.50, p=0.17 for corn; Wilcoxon test= 175.00, p=0.97 for sunflower ). Damage by monk parakeets was observed on 29 of the 30 sampled fields with sunflower (020 % of the plants damaged), and in 12 of the 24 corn fields (0 11% of plants damaged). Monk parakeet abundance was strongly correlated with damage in corn fields (R=0.75, p< 0.001), but not in sunflower fields (R=0.49, p= 0.01). Monk Parakeet Abundance and D amage in Crop Fields in Relation t o Within field, Field a nd Landscape Variables I did not detect any association of withinfield, field or landscape characteristics with abundance of monk parakeets in corn and sunflower fields. Top performing models includ ed only one variable at all levels, and all variables produced models with similar AICc values for all univariate models Table 2 1 ). Additionally, all 95% confident intervals (CI) for coefficients for each of the predictor variables included zero, indicating these factors probably had no effect on monk parakeet damage i n corn or sunflower fields ( Table 2 1 ). Given the lack of explanatory power of all variables, I did not explore multi level models with abundance data. In contrast to abundance, parakeet damage to crop fields clearly was associated with within field, field, and landscape characteristics. Most variables representing withinfield and field characteristics were included in the top performing models at each level for either corn or sunflower (Table 2 2). Similarly, most landscape variables were

PAGE 33

33 included in the top performing models at each buffer extent (1, 3 and 5km, Table 2). However, differences emerged in the way explanatory variables were related to parakeet damage on corn and sunflower fields at each level. Within field and field level variables. Monk parakeet damage to corn and sunflower fields increased as weed coverage increased, and damage decreased as plant density increased (Table 2 2). Also, monk parakeet damage to corn fields decreased as phenological stage of corn advanced (Table 2 2). Plant density and phenological stage were the most important variables for explaining monk parakeet damage to sunflower and corn fields, r espectively, while weed coverage was less important for both crop type s (Table 2 3). At the field leve l, monk parakeet damage to corn fields decreased as the field shape become more irregular or different from a regular square (i.e., shape index increased) Monk parakeet damage to sunflower fields declined as field area increased (Table 2 2), and damage to sunflower fields increased as tree abundance on the field perimeter increased (Table 2 2). Field shape was the most important variable explaining monk parakeet damage to corn fields, and field area and tree abundance were the most important variables expl aining monk parakeet damage to sunflower fields (Table 2 3). Landscape level variables. Monk parakeet damage to corn fields was related positively to the percentage of landscape with trees and pastures and other agricultural uses (including weedy and fallow fields) around the field at multiple buffer extents (Table 2 2). Additionally, monk parakeet damage to corn fields increased where crops susceptible to damage (corn and sunflower) by monk parakeets were more aggregated on the landscape. In sunflower fiel ds, the relationship between monk parakeet damage

PAGE 34

34 and aggregation of crops susceptible to damage was less clear (Table 2 2). Parakeet damage to sunflower fields increased as distance to the nearest site including man made structures and trees declined (Table 2 2). For both crop types, t he percentage of landscape with trees consistently had the lowest AICc value and was among the most important variables explaining monk parakeet damage to crop fields at multiple buffer extents (Table 2 3). The damage of monk parakeets to crop fields was better explained by models including variables at landscape level than for models describing the field characteristics or conditions within the field. Based on AICc values, single level model s with variables within 1 km and 3km buffers performed better than any other singlelevel models for corn and sunflower fields, respectively ( Table 2 2 ). For corn fields, the singlelevel model with variables within 1km buffer performed even better than the multi level model i n explaining monk parakeet damage to crop field s (Table 2 2). However, for sunflower fields, the multi level model, including a landscape variable in addition to within field and field variables, performed better than any of the singlelevel models ( Table 2 2 ). Discussion Factors R elated with M onk Parakeet Abundance and D amage to C rop F ields Monk parakeet abundance and damage were greater in sunflower than in corn fields. Although abundance and damage were correlated, I could explain monk parakeet damage based on within field, field and landscape variables but I could not explain monk parakeet abundance i n field s based on any variable that I measured. Probably, because bird damage is cumulative in time and bird abundance is not (Hone 1994), bird damage d ata allowed me to better capture differences among fields in relation to within-

PAGE 35

35 field, field and landscape variables than abundance data. Additional statistical analyses that I conducted with abundance data using Classification and Regression Trees (CART) showed relation ships similar to the se reported here for damage data for some variables, such as an increase in monk parakeet abundance in sunflower fields as plant density decreased or percentage of landscape comprised of tree patches around the field increased. However the performance of models in explaining monk parakeet abundance was g enerally poor based on the low percentage of variance explained by models at each level (usually between 30 and 50%). Landscape variables were more important than local variables in ex plaining monk parakeet damage to corn and sunflower fields. Particularly, the distance to nearest sites with trees and percentage of the landscape with tree patches around the crop fields were consistently important for explaining monk parakeet damage to c orn or sunflower fields. The increase in damage in sunflower fields proximate to sites with manmade structures and trees, and in corn or sunflower fields surrounded by abundant trees, may reflect lower energetic costs for monk parakeets traveling short di stances from the nest or loafing areas to foraging sites. This pattern may indicate the importance of landscape processes, such as landscape complementation, for a central place forager such as the monk parakeet. Landscape complementation refers to the occ urrence of habitat patches containing nonsubstitutable resources for a species in close proximity ( i.e., food and nesting sites, Dunning et al. 1992). Consequently, species abundance in a particular patch would be larger if the patch is located in a lands cape in which both habitats are relatively common or in close proximity rather than in a landscape in which one habitat is rare or both habitats are far apart (Dunning et al. 1992). Considering monk parakeets

PAGE 36

36 use nests all year around, both for breeding a nd roosting, and they travel limited distances each day from the nest to foraging sites (Spreyer and Bucher 1998), the abundance and/or proximity of tree patches with potential nesting sites could influence population size of parakeets and, consequently, damage on particular plots within those landscapes. This result supports the finding of a study of red winged blackbirds in which the availability of roosts and loafing areas in proximity to sunflower fields (2.4 km) was important for explaining differenc es on bird damage among fields (Hagy et al. 2008). The availability of alternative foraging sites for monk parakeets within the landscape also was important for explaining monk parakeet damage to corn and sunflower fields. Crop damage was positively assoc iated with the degree of aggregation of fields containing preferred food crops in the landscape and with the percentage of the landscape with pastures and other agricultural uses that could provi de weed seeds and other foods. This pattern may indicate that parakeets are spending more time foraging in fields where other foraging sites are easily available, or that populations are larger in areas with alternative foraging sites. This result may indicate the importance of another landscape process, landscape s upplementation, for monk parakeets. Landscape supplementation occurs when patches with substitutable resources occur in proximity in the landscape and, therefore, sustain a larger population than does a landscape in which these habitats are far apart (Dunning et al. 1992) Considering monk parakeets have a generalist forag ing behavior (Bucher et al. 1991 ; Hyman and Pruett Jones 1995), we w ould expect the availability of alternative foods on the landscape to allow potentially higher populations in the area a nd, consequently, higher damage in

PAGE 37

37 particular plots within that area. However, this result differs from studies of other bird pest species in which the availability of alternative foraging sites around crop fields was negatively related to bird use or dam age on those fields (Amano et al. 2004, 2008; Hagy et al. 2008). The high preference of monk parakeet for sunflower and, at lesser extent, corn compared to other seeds (Arambur and Bucher 1999; Canavelli, unpublished) may explain damage in these fields, even when alternative seeds were available in the landscape. Although landscape characteristics around the crop fields we re important for predicting monk parakeet damage to crop fields, local variables (withinfield and field level) also we re important Fo r corn fields, the characteristics of the landscape within 1 km of the field we re more important than any local characteristic for explaining parakeet damage. However, phenological stage of cr op plants and field shape also were important at local levels w ith greater m onk parakeet damage i n immature and regular ly shaped corn fields. F or sunflower fields, local characteristics of the field, such as plant density and field area, we re as important as landscape variables Small sunflower fields, with low plant density and high percentage of the field border with trees we re more prone to monk parakeet damage than other fields. These results support the need to consider landscapelevel variables and local level variables for predicting bird damage to crop fields ( Tourenq et al. 2001; Amano et al. 2008; Hagy et al. 2008). Differences on the degree of importance of local variables compared to landscape variables for corn and sunflower could be explained by differences i n structur al characteristics of each crop on the ground C orn is known to have less vegetative and reproductive plasticity than sunflower under different levels of plant density (Oscar

PAGE 38

38 Valentinuz, pers.com.) Consequently, s mall variations in plant density hav e lower influence on the structure of indiv idual plants in corn compared to sunflower (Oscar Valentinuz, pers.com.) As a result, corn fields usually a re characterized by a more homogenous stand of plants compared to sunflower fields (i.e., individual plants in corn usually are more similar among each other compared to sunflower plants) The high variation in fine scale structure of sunflower fields could provide the basis for patch selection by birds at fin e r scales and could help to explain the higher importance of local variables for sunflower fi elds compared to corn fields. The use of multiple buffer extents have been shown to improve the detection of birds responses to landscape context (e.g., Cooper and Walters 2002; Renfrew and Ribic 2008; Boscolo and Metzger 2009). In this study, models at t he buffer width of 1 km for corn fields and 1 and 3 km for sunflower fields had better performance than models at the buffer width of 5 km. Assuming these buffer extents reflect the scale of the foraging process under study, the results are consistent with reduced mobility and small home range of monk parakeets during nesting season (Bucher 1985, 1992b) and may indicate that daily distances of travel between nest and foraging sites are even shorter than originally proposed ( i.e., between 3 and 5 km, Spreyer and Bucher 1998). Short distances of travel could be related to r eproductive characteristics of monk parakeets considering most crop fiel ds were sampled at the end of the reproductive season (January) and v ariables related to potential places for nesting such as the distance from the crop fields to the nearest site with manmade structures and trees and the percentage of the landscape occupied by tree patches were among the most important factors explaining monk parakeet damage at all buffer extents. Ho wever, t he importance

PAGE 39

39 of variables within 13 km buffers around fields also may be related to the feeding characteristics of parakeets because, given their generalist and omnivorous diet, they could easily shift among food items within a relatively small area (Boscolo and Metzger 2009). Management I mplications Monk parakeet damage to corn and sunflower fields is an important concern for agricultural producers in some areas of Argentina ( Bucher 1984, 1992a, b; Bruggers y Zaccagnini 1994; Bruggers et al. 1998 ) Although damage values in this study were relatively low ( < 5 % of damaged plants, corresponding to less than 3% of grain loss, Canavelli et al. 2008), farmers perceive a problem exists, and sometimes apply short term control methods, such as toxic bait s and nest poisoning with pesticides (De Grazio and Besser 1975; Arambur 1991; Bucher 1992 a,b ) However, these management measures have low success to decrease damage on specific crop fields (Bruggers et al. 1998; Spreyer and Bucher 1998; Ro driguez and Zaccagnini 1998) and potentially have severe consequences for nontarget wildlife species (Zaccagnini 2006). For this reason, objections to these types of control have occurred in the past and are currently increasing, and new methods are needed (Bucher 1992 a b ; Bruggers et al. 1998; Canavelli and Aramburu, in press ). In order to better plan integrated management schemes for preventing and/or decreasing monk parakeet damage, wildlife managers need information on the underlying factors that promote the damage of monk parakeets on specific crop fields Crop type and landscape context are very important for explaining monk parakeet damage to corn and sunflower fields. Monk parakeet damage was significantly more frequent and higher in magnitude in sunflower than i n corn fields. Therefore, farmers

PAGE 40

40 planning to plant sunflower should be more aware of monk parakeet damage and, consequently, plan management alternatives to decrease monk parakeet damage more carefully than farmers planting corn. Also, in order to prevent monk parakeet damage, crop fields should be located at least 1 km from places with manmade structures and trees (cascos) and other patches with trees, which could be particularly difficult in some areas without cutting down the trees. Additionally, far mers planting sunflower fields could reduce damage by increasing field area. Both eliminating trees and increasing field area, and consequently decreasing the availability of alternative noncrop habitats, such as arborous or weedy edges, would result i n a simplified landscape, which could have important negative consequences not only for biodiversity b ut also for the regulation of other crop pests (Bianchi et al. 2006; Power 2010; Batry et al. 2011). Therefore, when the magnitude of damage justifies apply ing management measures, these alternatives should be considered with caution and evaluated in relation to other strategies, such as crop protection, in the context of a regional management strategy. Finally, farmers may consider increasing plant density as a measure to prevent monk parakeet damage. In this case, farmers also should consider the tradeoff of this agronomical practice and the limits on plant density recommended for each crop in the area, in order to avoid other problems such as smaller plants and/or yields. Alternative feeding areas or lure crops probably would not be successful in decreasing damage by monk parakeets in particular crop fields. However, this management al ternative needs more evaluation particularly with sunflower as the attrac tive crop.

PAGE 41

41 Figure 2-1. Map showing the location of Department of Paran (Entre Ros Province, Argentina) and the crop fields sampled in 2007 and 2008. Black dots indicate corn fields (n=25) and grey dots indicate sunflower fields (n=31).

PAGE 42

42 Section 1 Section 2 Section 3 Section 4 Sampling row Additional short transects in field border (2008) Plants Field edge Field border Field center Crop row Plants Section 1 Section 2 Section 3 Section 4 Sampling row Additional short transects in field border (2008) Plants Field edge Field border Field center Crop row Field edge Field border Field center Crop row Plants Figure 2 2. Sampling scheme for evaluating bird damage to crop fields.

PAGE 43

43 Figure 2 3. Monk parakeet damage to corn ears. Figure 2 4. Monk parakeet damage to sunflower heads.

PAGE 44

44 Figure 25. Buffer extents used for sampling landscape level variables from satellite images around each crop field (central point) 1 000 m 3000 m 5000 m

PAGE 45

45 Table 2 1 Minimum AICc models and regression results for factors used in predicting monk parakeet abundance in crop fields in Paran depart ment (Entre Ros, Argentina) in 2007 and 2008 summer seasons. All models are within each other Multi level models were not run because I could not clearly identify the most important variable within each level Coefficients and associated standard errors for each predictor variable are derived from the univariate model and the Akaike weight ( i). I n landscape models, numbers preceding the variable code indicate buffer sizes (in kilometers) Corn Sunflower Spatial level Variable 1 AICc Akaike weight i ) Coefficient CI AICc Akaike weight i ) Coefficient CI Within field PLTDEN 25.27 0.27 0.15 0.60 35.63 0.27 0.01 0.29 PHENST 25.36 0.26 0.45 0.90 35.64 0.27 0.08 0.39 WDCOV 25.36 0.26 0.96 2.06 35.65 0.27 0.43 1.27 Field AREA 25.23 0.27 0.01 0.05 35.63 0.27 0.003 0.02 SHAPE 25.25 0.26 1.79 2.08 35.63 0.26 0. 45 1.57 TREES 25.33 0.25 0.62 1.47 35.66 0.26 0.05 0.40 Landscape 1 km DISTCO 25.35 0.14 0.002 0.004 35.64 0.14 0.0002 0.002 1CRPLAN 25.23 0.15 0.01 0.11 35.69 0.14 0.02 0.03 1CRCLUMP 25.24 0.15 1.94 5.73 35.66 0.14 1.21 3.87 1TRPLAND 25.29 0.14 0.02 0.06 35.66 0.14 0.006 0.02 1PSTPLAND 25.30 0.14 0.03 0.08 35.71 0.14 0.02 0.04 Landscape 3 km 3CRPLAN 25.27 0.19 0.03 0.13 35.81 0.17 0.03 0.04 3CRCLUMP 25.20 0.20 3.40 7.67 35.58 0.19 0.90 3.35 3TRPLAND 25.48 0.17 0.03 0.04 35.59 0.18 0. 003 0.01 3PSTPLAND 25.21 0.19 0.03 0.12 35.58 0.19 0.001 0.05 Landscape 5 km 5CRPLAN 25.26 0.19 0.03 0.14 35.71 0.18 0.02 0.04 5CRCLUMP 25.13 0.20 2.70 3.20 35.70 0.18 3.12 6.27 5TRPLAND 25.53 0.16 0.03 0.04 35.64 0.18 0.002 0.01 5PSTPLAND 25. 24 0.19 0.16 0.06 35.64 0.18 0.01 0.05 1 PLTDEN = Plant density, PHENST = Phenological stage, WDCOV = Weed coverage, AREA = Field area, SHAPE = F ield shape index, TREES = Abundance of trees on border, DISTCO = D istance to the nearest site with man made s tructures and trees, CRPLAND = Percentage of the landscape with crops susceptible to damage (corn and sunflower), CRCLUMP = Clumpiness index of crop patches susceptible to damage, TRPLAND = Percentage of the landscape with tree patches, PSTPLAND = Percenta ge of the landscape with pastures and other agricultural uses (including weedy and fallow fields).

PAGE 46

46 Table 2 2 Minimum AIC c mo dels for monk parakeet damage to c orn and sunflower fields in Entre Rios (Argentina) during 2007 and 2008 summer seasons. within each level Models are ordered based on model performance within each level (lower AICc value indicating better model performance). Brackets indicate a negative relationship with bird abundance or damage Variables are defined in Table 1. I n landscape and m ultilevel models, numbers preceding the letters code indicate buffer sizes (in kilometers) Corn Sunflower Spatial level Model AICc Akaike weight i ) Model AICc Akaike weight i ) Within field ( PHENST) ( PHENST) + WDCOV ( PHENST) + ( PLTDEN) ( PHENST) + ( PLTDEN)+ WDCOV ( PLTDEN)+ WDCOV ( PLTDEN) 63.90 64.17 64.26 64.71 65.85 65.88 0.24 0.21 0.20 0.16 0.09 0.09 ( PLTDEN) + WDCOV ( PLTDEN) 306.40 307.10 0.46 0.32 Field ( SHAPE) 63.37 0.62 ( AREA) TREES 315.59 317.14 0.63 0.29 Landscape 1 km 1TRPLAND + 1PSTPLAND +1CRCLUMP 22.70 0.91 ( DISTCO) + 1TRPLAND ( DISTCO) + 1TRPLAND + 1CRCLUMP 281.48 282.84 0.44 0.22 Landscape 3 km 3TRPLAND + 3PSTPLAND + 3CRCLUMP 3TRPLAND + 3CRC LUMP 3TRPLAND + 3PSTPLAND 35.40 35.95 36.47 0.41 0.31 0.24 3TRPLAND + 3PSTPLAND 276.68 0.77 Landscape 5 km 5TRPLAND + 5PSTPLAND + 5CRCLUMP 5TRPLAND + 5PSTPLAND 42.49 43.79 0.60 0.31 5TRPLAND + 5PSTPLAND 5TRPLAND + 5PSTPLAND + ( 5CRCLUMP) 287.12 288.13 0.62 0.38 Multi level ( PHENST) + ( SHAPE) + 1TRPLAND 38.93 0.96 ( PLTDEN) + ( AREA) + 1TRPLAND 249.34 0.99

PAGE 47

47 Table 2 3 Regres sion results for factors considered in predicting monk parakeet damage to crop fields in Paran department (Entre Ros Argentina) in 2007 and 2008 summer seasons. Coefficients and associated standard errors for each predictor variable are derived from multimodel inferences using all parameter subsets and Akaike weights ( i) at each level i for each predictor variable shows the sums of Akaike weights for all important a variable is relative to other variables. Corn Sunflower Spatial level Variable AICc Coeff icient CI i AICc Coefficient CI i Within field PLTDEN 65.88 0.17 0.18 0.54 307.10 0.27 0. 1 0* 1 00 PHENST 63.90 0.51 0.40* 0.81 333.23 0.005 0.03 0.22 WDCOV 69.52 0.2 7 0.3 1 0.4 7 331.50 0.1 7 0.1 9 0. 44 Field AREA 71.63 <0.001 <0.001 0.05 320.72 0.01 0.00 8 1.00 SHAPE 63.37 3 14 2 70 0. 83 328.79 0.11 0. 20 0. 32 TREES 71.63 0.008 0.09 0.18 329.86 0.17 0.12* 0. 93 Landscape 1 km DISTCO 70.40 <0.001 <0.001 0.01 326.83 ( ) <0.001 <0.001* 0. 66 1CRPLAN 69.97 <0.001 <0.001 0.00 324.64 <0.001 < 0.001 0.00 1CRCLUMP 55.10 13. 77 11. 58 0.9 7 325.50 0.2 7 0. 51 0. 31 1TRPLAND 33.79 0.06 0.03* 1 00 283.58 0.0 2 0.00 6 1.00 1PSTPLAND 69.66 0.06 0.04* 0.9 3 325.90 0.002 0.002 0.17 Landscape 3 km 3CRPLAN 48.54 0.001 <0.001* 0.01 308.59 ( ) <0.001 <0. 001 0.00 3CRCLUMP 56.13 9. 73 9.8 9 0.7 3 304.19 0.18 0.64 0.22 3TRPLAND 40.72 0.05 0.02* 0.9 9 334.17 0.02 0.00 6 1.00 3PSTPLAND 71.66 0.06 0.06 0.65 323.65 0.0 4 0.01* 1.00 Landscape 5 km 5CRPLAN 52.31 0.02 0.01* 0.07 319.06 ( ) <0.001 <0.001* 0.00 5CRCLUMP 70.83 1 2 61 1 5 58 0.64 332.13 0.63 0.92 0.38 5TRPLAND 51.24 0.0 6 0.0 2 0.93 317.25 0.02 0.006* 1.00 5PSTPLAND 71.59 0. 14 0.0 8 0.98 320.16 0.05 0.02* 1.00 95% confidence intervals for multimodel weighted coefficients for each of the pre dictor variables not including zero (i.e., these factors probably affected monk parakeet damage to corn or sunflower fields).

PAGE 48

48 CHA P TER 3 DENSITY OF MONK PARA KEET NESTS IN SITES WITH EUCALYPTUS IN RELATION TO LOCAL AN D LANDSCAPE VARIABLES Background The se lection of nesting sites by birds is one of the most critical processes involved in species persistence, because of its direct influence on reproductive success and population growth (Walsberg 1981; Li y Martin 1991; White et al. 2006). Similar to other habitat selection processes, selection of nesting sites involves a hierarchy of decisions at multiple scales (Hildn 1965; Johnson 1980; Cody 1985; Wiens et al. 1987). Landscape characteristics may influence initial settlement in a site or patch ( area of con tiguous cover different from its surroundings, Forman and Godron 1986), and local characteristics of the site, such as vegetation structure, influence selection of a particular place for the nest (Hildn 1965; Bailey and Thompson 2007). Multi level studies in which researchers evaluate the influence of local and landscape variables on animal abundance and distribution in particular patches ( i.e., focal patches, Brennan et al. 2002) have been used for understanding factors influencing nest site selection by birds (Donzar et al. 1993; Soh et al. 2002; Martinez et al. 2003; Bailey and Thompson 2007). Multi level studies are common in landscape ecology and conservation biology (see reviews in Mazerolle and Villard 1999 and Thorton et al. 2010). Also, multi level studies have been proposed as a very helpful tool for integrating multiple scales of observation a bout the behavior of birds involved in humanwildlife conflicts and designing management strategies (Clergeau 1995; Soh et al. 2002). In this study, I applied a mul ti level approach for analyzing local and landscape factors influencing density of monk parakeet nests in sites with e ucalyptus trees in east

PAGE 49

49 central Argentina. The monk parakeet ( Myiopsitta monachus ) is a parrot species commonly involved in humanwildlife conflicts in its native range (South America) and nonnative areas of distribution (North America and Europe, Spreyer and Bucher 1998). In Argentina and Uruguay, the monk parakeet is among the most important bird pest species causing damage t o grain crops (Bucher and Bedano 1976; Bucher 1984, 1992a, b; Bruggers y Zaccagnini 1994; Bruggers et al. 1998). This species also causes damage in other settings (e.g., fruit crops and electric utility structures, Bucher 1992a; Bucher and Martin 1987). In North America and Europe, the monk parakeet primarily causes problems in urban settings related to location of nests, including damage to electric utility structures that lead to power outages (Avery et al. 2002) and disturbance to tranquility of human neighborhoods, due to the noise parakeets produce (Santos 2005; Burger and Gochfeld 2009). Currently, the monk parakeet is not a threat to agriculture production or native biota in nonnative areas, but if its populations continue to expand, this species could cause more problems in the future (Sol et al. 1997; Domenech et al. 2003) Understanding habitat features influencing the selection of nesting sites by monk parakeets is important for managing conflicts with this species in its native an d non native areas of distribution. Most studies on nesting habitats for monk parakeet have focused on describing substrates selected for nesting and characteristics of nests (Humphrey and Peterson 1978; Sol et al. 1997 ; Burger and Gochfeld 2000, 2005, 2009). S uitable nes ting sites for monk parakeets are varied because unlike other parrots, the monk parakeet does not nest in cavities but rather constructs nests with sticks on tall natural and artificial structures (Forshaw 1989; Spreyer and Bucher 1998). Natural structures include

PAGE 50

50 savanna trees (e.g., Prosopis sp. and Acacia sp.), willow ( Salix sp.), palms (e.g. Phoenix sp.), and eucalyptus ( Eucalyptus sp.), among others (Sol et al. 1997; Spreyer and Bucher 1998; Gochfer and Burger 2000, 2005, 2009). Artificial structures for nesting include silos, fire escapes, windmills, and utility poles (Spreyer and Bucher 1998). In spite of the extensive information about the characteristics of substrates selected for nest placement almost no information is available about the influe nce of local and landscape habitat features on the abundance of monk parakeet nest s in a specific site or patch in its native range (but see Cavallero 2010). In Spain, selection of nesting patches by monk parakeets was explained by microhabitat variables, particularly the type and height of the trees, and macrohabitat variables, such as abundance of palm trees, a preferred nesting substrate in that area (Sol et al. 1997). In the Pantanal of Brazil, selection of nesting patches with colonies of monk parakeets was influenced for pre existing nests of jabiru stork ( Jabiru mycteria) and the availability of large trees, usually planted in proximity to farm houses ( Burger and Gochfeld 2005). In Argentina, the only study that has examined local and landscape vari ables that influence the abundance of monk parakeet nests was done in patches with native savanna trees, such as Prosopis s p. and Acacia sp. (Cavallero 2010). In that study, the relationships between abundance of monk parakeet nests and habitat characteris tics of patches were found to be weak at both local and landscape levels, though local characteristics, such as tree density and cover, were slightly more important than landscape characteristics (Cavallero 2010).

PAGE 51

51 In Argentina, monk parakeets apparently prefer introduced eucalyptus trees for nesting because they are taller than the surrounding native trees, which could reduce the i ncidence of nest failure from predation, particularly from humans ( Humphrey and Peterson 1978 ; Navarro et al. 1992 ) In agricul tural landscapes throughout central Argentina, eucalyptus trees are usually planted around manmade structures (e.g., houses and barns) for shade and windbreaks. These areas of human habitation with eucalyptus trees are predominantly common as nesting areas for monk parakeets in this highly modified landscape (Spreyer and Bucher 1998). Additionally, these areas of human habitation with eucalyptus trees are sites where monk parakeets potentially are subject to control measures. Therefore, I focused this st udy on these types of sites (referred to as farms hereafter). I hypothesized that the density of monk parakeet nests in a site would be related to two factors in the patch: 1) e ucalyptus trees, specifically abundance and height of these trees, and 2) whether control measures for parakeets were applied in the patch. I also expected characteristics of landscape around the patch, such as the abundance of nesting and foraging habitats, to influence the density of monk parakeet nests in the patch. However, based on other studies of nests in native habitat, I expected landscape characteristics to be less important than local characteristics for explaining density of monk parakeet nests in the patch. Methods Study A rea The study was conducted in a 525,000ha area c omprising the Department of Paran (Entre Ros Province, Argentina). The area is characterized by diverse production activities, with predominance of crops, beef cattle, and milk production (Engler and Vicente 2009). Agricultural crops cover about 49% of t he area, including in

PAGE 52

52 order of importance soybeans, wheat, corn, sunflower and sorghum (Engler and Vicente 2009). Mean annual temperature is 19C (12C in winter and 25C in summer) and mean rainfall is approximately 1000 mm. A gradient in production acti vities, and therefore landscape pattern, occurs in the study area from northeast to southwest. The northeast mostly is devoted to a mixture of crops and cattle production, with high interspersion of woodlands, pastures and crops. The southwest is intensively agricultural, with a few patches of woodland interspersed with large plots of annual crops. Sampling S cheme I used a Geographic Information System (ArcGIS v.9.2) to place a 10x10km grid over the Paran Department and identified the most proximate site with eucalyptus trees to the center point in each grid. This site constituted the first sampling patch in each grid cell. A patch was defined as an area of at least 900 m2 (30x30 m) including eucalyptus trees clearly differentiated from its surroundings and located at least 30 m from another patch. A second patch was selected in the field at least 2000 m from the previous sampling site, which represents the estimated maximum distance for dispersal by monk parakeets ( Spreyer and Bucher 1998) to assure independence of birds among nesting sites When eucalyptus trees were seen on the horizon from the first sampling patch, they were used as a reference for searching for the second sampling patch. Otherwise, the second sample was located by following availa ble routes around the central point in the grid cell. Between September and December 2008, I identified 62 sites with e ucalyptus trees around occupied human habitations and other manmade structures (e.g., barns, water tanks) I used these 62 sites for this study.

PAGE 53

53 Nest Abundance S urveys Monk parakeets build communal nests, comprising many nests in one nest structure, usually grouped in colonies with many nest structures in one area (Forshaw 1989; Bucher et al. 1991; Eberhard 1998; Burger and Gochfeld 2005, 2009). For the purpose of this study, a nest was defined as a distinct and compact collection of twigs with one or more nest cavities or chambers Nest size as well as distance between nests in a colony is highly variable. For example, nest size can rang e between < 0.80 m to > 1.5 m in diameter (Spreyer and Bucher 1998), and distance to the nearest nest can range from 8.54 to 60.20 m (Burger and Gochfeld 2000). However because nests are very compact, separate nests are easy to distinguish (Forshaw 1989; Spreyer and Bucher 1998; Burger and Gochfeld. 2009). Additionally, nest s usually are placed on the uppermost branches of tree (Forshaw 1989; Eberhard 1998; Spreyer and Bucher 1998although see Burger and Gochfeld 2005), which makes them easy to detect from the ground. All nests in a patch were considered part of the same colony. Two observers systematically surveyed each patch by walking along lines with e ucalyptus trees, one to each side of the tree line, counting al l parakeet nests seen in the t rees. When isolated trees were found, the tree was circled and all nests observed i n the tree were recorded. Nests i n introduced tree species other than eucalyptus (2 cases) and humanmade structures (1 case) only were observed on abandoned farms, and therefore were not included in the study. In addition to the number of nests, the number of nest cavities or chambers per nest also was recorded. However, because of the difficulties of accurately determini ng the number of nest cavities i n each nest from the ground, and the high correlation between abundance of nests and nest cavities in nesting patches (r=0.92), I considered only the abundance of nests in this study.

PAGE 54

54 Local L evel V ariables Local data for each farm included three measures of eucalyptus trees: 1) height of trees, taken as the mean of the tallest and shortest tree in the patch, 2) canopy area, and 3) proportion of the patch area occupied by eucalyptus canopy. Canopy area was estimated by projecting the outer edge of the canopy to the ground and then measuring the area of the polygon delimited by the outer edge in the field (Hays et al. 1981). In the case of individual eucalyptus trees, the canopy diameter was measured with a measuring tape and the area of the polygon was estimated based on a circle area. W hen eucalyptus trees were in a line, the area of the polygon was estimated by measuring the length and width of the polygon with a tape and assuming the polygon was a rectangle. Additionally, I recorded presence/absence of control measures applied in the s ite against monk parakeets and the type of control being applied by interviewing the landowner. Finally, I recorded the geographic coordinates of patch borders in the field. Once in the lab, all patches were digitized using Google Earth and the geographic coordinates of patch borders. Polygons were then converted to a vector file and imported in ArcGIS (v 9.2) and the patch area was estimated using Patch Analyst extension for ArcGIS (Rempel 2010) Because s ubstantial correlation ( r=0.70) occurred between patch area and canopy area of eucalyptus in the patch and only uncorrelated metrics were included in the same statistical model, I used canopy area of eucalyptus for modeling purposes (Statistical analysis), based on its significance as a nesting substrate for monk parakeets.

PAGE 55

55 LandscapeS cale Variables I evaluated landscape context around each nesting patch within circular buffers of 3 extents (1000, 2000 and 3000 m) from the center of each patch ( Figure 3 1) The upper buffer limit of 3000 m was greater than the documented maximum dispersal distance for monk parakeets (2000 m, Speyer and Bucher 1998). In some cases 3 km buffers of contiguous nesting patches exhibited > 50% overlap and, i n these cases, I selected one of the p atches for statistical analys is based on its proximity to the central point on the original 10x10 grid. I did not use buffers smaller than 1000 m because of problems with artificial borders and estimation of landscape indices. Buffers for nesting patches were obtained from a Landsat TM image classified by Noelia Calamari (INTA, EEA Paran). The Landsat image was classified using supervised classification with ECHO (Extraction and Classification of Homogeneous Objects) in MultiSpect Application v3.1 (2007). Ten land cover types were identified: water, corn, sunflower, soybeans, sorghum, pastures and other agricultural uses (e.g., fallow and weedy fields), built up areas, plowed and some fallow fields, which could not be clearly distinguished, introduced trees, native trees, and riparian vegetation. Results from the classification were validated with 100 points per land cover type randomly selected using Quickbird images (available in GoogleEarthTM, http://earth.google.com ) and ground sampling. Overa ll classification accuracy was 82 % (range for cover types: 54% for corn and 96% for water ) Finally to clearly distinguish among different land cover types and decrease the problem of artificial borders with raster images, I regrouped the land cover ty pes into 7 classes: water, preferred food crops for monk parakeets (corn and sunflower), unus ed crops (soybean and sorghum, immature at the

PAGE 56

56 time the image was obtained), pastures and other agricultural uses (e.g. fallow and weedy fields), built up areas, plowed and fallow fields, and tree patches including predominantly native and, at much les ser extent, introduced trees. In this analysis, I used only availability of three land cover classes: 1) preferred food crops (corn and sunflower), 2) tree patches (all patches on the landscape with native or introduced trees), potentially used as primary sites for perching, nesting or daily loafing, and 3) pastures and other agricultural uses (e.g., fallow and weedy fields), which can include food items for monk parak eets such as flowers and seeds. I used FRAGSTATS 3.3 software (McGarigal et al. 2002) to calculate landscape metrics representing landscape composition and configuration. Composition metrics included percentage of landscape with corn and sunflower, other agricultural uses (including pastures, fallow and weedy fields) and tree patches. Configuration metrics were estimated for tree patches and included patch density, edge density for tree/nontree patches, mean nearest neighbor distance among tree patches and aggregation of tree patches (measured with a clumpiness index) Shape complexity was measured using a shape index, estimated as the patch perimeter (m) divided by the square root of patch area (m2), adjusted by a constant to adjust for a circular stand ard ( McGarigal et al. 2002) Substantial correlations ( r 0.60, independently of the pvalue) were found among some landscape metrics, particularly at higher extents (2000 and 3 000 m Table A 3, Appendix A ) and only uncorrelated metrics ( r<0.60) were included in the same statistical model. Because patch density was the only configuration metric representing fragmentation of tree patches that was uncorrelated with percentage of landscape with trees, and percentage of landscape with trees could be important for explaining bird

PAGE 57

57 abundance at a site ( Fahrig 2001 200 3 ; Renfrew and Ribic 2008; Hagy et al. 2008), I only included patch density as a configuration metric in statistical models Percentage of landscape with pastures and other agricultural uses was negatively correlated with percentage of landscape w ith tree patches particularly at higher extents ( 2000 and 3 000 m Table A 3, Appendix A ). Because tree patches represent alternative nesting sites and pastures and other agricultural uses represent potential foraging sites, I retained both variables but included only one of them at a time for model construction. Statistical A nalyses I modeled density of monk parakeet nests in nesting patch as a function of local and landscape variables at each buffer extent (1000, 2000 and 3000 m). Additionally, I constructed a set of models combining local and landscape variables. The response variable (density of monk parakeet nests) was examined for deviation from normality and transformed to the decimal logarithm to meet normality and improve linearity (Infostat, Di Rienzo et al 2010; SAS v. 8.0, SAS Institute Inc., 2006). D ata from one nesting patch w as consistently outside the distribution range of values for all independent variables and w as eliminated from the pool of data, as well as five patches with missing values for at least one variable at the local level Additionally, twenty patches were eliminated for the landscape analysis, because of overlapping buffers. Therefore, final sample size for modeling purposes was 35 nesting patches though descriptive statistics are presented for all patches (Table 31) All models were developed using a generalized linear model framework (GLM) with a normal error structure. Because final sample size for model construction was relatively small (n=35), each model was restricted to incl ude between one and three

PAGE 58

58 variables. I ran models for all single variables, all sets of two variables and then all sets of 3 variables within each level (local and landscape). Then, I ran multi level models that contained the strongest predictor from local and landscape levels. For multi level models, I used the local variable with the minimum AICc value as the base model and then added the landscape variable with the best performance at each buffer extent (Fletcher and Koford 2002, Renfrew and Ribic 2008). All models were evaluated using SAS PROC GENMOD and maximum likelihood estimation. I used Akaike Information Criteria adjusted for small sample size (AICc) for comparing model performance, and I restricted AICc scores to 4 for model retention (Burham and Anderson 2002). For evaluating individual variable performance at each level, I used model averaging and the sum of plausible models in which a variable was i, Burham & Anderson 2002). Results Density of P arakeet N ests and C haracteristics of N est Patches and L andscapes Most surveyed patches (85%) had monk parakeet nests in the eucalyptus trees. However, density of nests among patches varied widely (range: 1222 nests/ha, mean= 27 nests/ha 5.56 SE ). Area covered by canopy of eucalyptus tre es in a patch generally was small ( < 1 ha), occupying o n average about a quarter of the patch area (Table 3 1). Mean height of eucalyptus trees was relatively uniform among patches (CV= 19%, Table 3 1). At most sites (63% ), people had not applied control measures against monk parakeets in the last 2 years. At patches where control measures occurred (n=23), most people applied lethal control by shooting, either as the only control measure (52%) or combined with removing (9%) or burning (17%) nests. Nonleth al measures included removing nests (4%), burning nests (9%), or capturing

PAGE 59

59 nestlings alive for pet trade (9%). Trees occupied on average a quarter of the landscapes surrounding nest patches, although a wide range of variation was observed (Table 3 1). Pref erred food crops for monk parakeets (corn and sunflower) and pastures and other agricultural uses were less abundant, although these habitats together could occupy more than a quarter of the landscape (Table 3 1). Density of M onk Parakeet N ests in Relation to Local and Landscape V ariables All local variables were included in the top performing model explaining density of monk parakeet nests at specific patches (Table 3 2 local level ). The density of nests was positively related to area of eucalyptus canopy proportion of the patch area covered by eucalyptus canopy, height of the trees, and presence of control measures for monk parakeets (Table 3 2 ) In the latter case, landowners were more likely to control parakeets when they were very abundant. However, the proportion of patch area with canopy of e ucalyptus trees was the only important variable for explaining density of monk parakeet nests at this level as indicated by the 95% confidence interval s for multimodel weighted coefficients for the predictor var iable s (Table 3 3). Similarly, all landscape variables were included in the top performing model explaining density of monk parakeet nests at specific patches (Table 3 2 landscape level at 1km, 2km and 3km ). Density of nests usually increased as the perc entage of the landscape with trees increased, density of tree patches in the landscape decreased (i.e., patches were more dispersed), percentage of the landscape with preferred food crops increased, and percentage of landscape with pastures and other agric ultural uses decreased (Table 3 2 and 3 3). At landscape level, the percentage of landscape with trees was the most important variable for predicting density of monk parakeet nests at all spatial extents (Table 3 3).

PAGE 60

60 Models generally had poor performance, with no model clearly the best for explaining variation in nest density among nesting patches at any level (Table 3 2). Models at single level that included the proportion of patch area with eucalyptus or the percentage of landscape with trees 1 km around the nesting patch had the lowest AICc values (Table 3 2). The best multilevel models at all spatial extents included these two variables and had lower AICc values than landscape level models, but not than local level models (Table 3 2). However, the degre e of uncertainty about a best model for explaining density of monk parakeet nests in farms with eucalyptus was high, because none of these models had a model weight greater than 0.36 (Table 3 2). Discussion Factors Related with D ensity of Monk P arakeet N es ts The density of monk parakeet nests in inhabited farms with e ucalyptus trees was not clearly explained by any variable or combination of variables evaluated in this study, either at local or landscape level. Although all variables were involved in the s ubset of top performing models, models performed poorly in explaining density of monk parakeet nests. These results are similar to results of a previous study conducted in the same region (Paran department and surroundings, Entre Ros, Argentina), in whic h the abundance of monk parakeet nests in patches with native trees was weakly related to local and landscape variables within 2.5 km of the nesting patch (Cavallero 2010). However, as in that study, some variables emerged as slightly more important at eac h level in this study, with local variables being comparatively more important than landscape variables The proportion of the patch area covered by eucalyptus canopy was the most important variable at the local level explaining abundance of nests. Addit io nally, this

PAGE 61

61 local variable had a better performance than any other variable at landscape scale (lower AICc value). T he canopy area of eucalyptus trees likely is an indicator of the availability of nesting sites for monk parakeets in a patch. Because monk p arakeets are colonial birds (Spreyer and Bucher 1998) and they prefer e ucalyptus trees for nesting ( Bucher and Martin 1987; Navarro et al. 1992) nest density would be expected to increase as the proportion of the patch occupied by eucalyptus increased. A t landscape level, the importance of percentage of landscape with trees around nesting patches in explaining the density of parakeet nests compared to other variables could be related to a variety of factors. For example, as the percentage of landscape wit h native tree patches increases, the abundance of monk parakeet s may increase because of the large amount of nesting habitat o r because patches of native trees contain some other resource that increases parakeet abundance (e.g., food). A relationship between the abundance of nest s in nesting patches and the abundance of trees around the patches, particularly preferred trees for nesting, also was found in previous studies with monk parakeets, both in its native and nonnative areas (Sol et al. 1997; Cavaller o 2010). The poor performance of local and landscape variables in explaining density of monk parakeet nests also could be related to several factors. S ome important variables may not have been measured or the way the variables were measured may not have been optimal for capturing the influence of those variables. For example, the type and structure of eucalyptus trees in each nesting patch which I did not measure, may explain differences in density of monk parakeet nests among patches. Tree type and struct ure have been found to be important i n explaining monk parakeet selection of tree

PAGE 62

62 species for nesting (Sol et al. 1997; Burger and Gochfeld 2005, 2009). Although all parakeet nests in this study were in eucalyptus trees, both p ersonal observations and a pr evious work (Volpe and Arambur 2010) indicate that nests a re usually placed in eucalyptus trees with stout trunks and open crowns that offered large and robust surfaces for nest construction. All these are characteristics of red eucalyptus ( Eucalyptus cam aldulensis or E.tereticornis Duke 1983), t he predominant type of eucalyptus tree I observed in the farms. However, other species of Eucalyptus trees freque ntly were available, including white eucalyptus, such as E.grandis or E. duniis, and E.cinerea. T he structure of the eucalyptus trees also may be more influenced by the arrangement and management of trees (isolated trees, trees in a line, or trees in small groups) than by the tree species (D.Diaz, pers.com.). Another issue may be that buffer extents us ed in this study were not large enough to capture habitat features related to nest settlement, such as availability of foraging sites I used the maximum dispersal distance to de fine the largest buffer extent but monk parakeets are know n to travel up to 8 km for daily foraging when breeding and 24 km when not breeding (Hyman and Pruett Jones 1995 ; Spreyer and Bucher 1998). Also, other landscape cover types, such as urban habitats, could be important for monk parakeets, but no data are available regarding t his. Alternatively, factors related to the behavior of monk parakeets could influence nest density more than habitat variables at local and landscape levels. Monk parakeets select a nesting site for the first nest, and then they continue adding nests to t he colony, building new nest chambers in the same nest or new nests in the proximity of the previous nest (Sol et al. 1997; Burger y Gochf eld 2000). F or colonial and philopatric

PAGE 63

63 birds, such as monk parakeets, the influence of landscape scale variables in nest site selectivity could be expected to be low er than for birds with no natal philopatry and high rate of colonization of new nest locations (Katie Sieving, pers.com.). Therefore, nest density may be related to differences in time of colonization of spec i fic patches instead of differences in habitat characteristics of the patches (Cavallero 2010) If this is case, the number or density of nests in nesting patch would be more dependent on the time the first nest appeared (longer ago, more nests) than to the characteristics of the patch. However if the best nest sites are chosen first, colony age and habitat quality could be linked. No information is available to elucidate this issue. The generalist foraging behavior of monk parakeets (F reeland 1973; Bucher et al. 1991; Hyman and Pruett Jones 1995 ; Spreyer and Bucher 1998), plus the abundance of food in the area from crops and other agricultural uses as well as weedy fields, may explain why the percentage of landscape with preferred food crops pastures and other agricultural uses contributed little to explaining density of monk parakeet nests. D ensity of monk parakeet nests in parks in Spain was higher in parks offering more plants for foraging, but the relationship between nest density and food plants was weak and less important than the abundance of nesting sites (Sol et al. 1997). A lso the abundance of monk parakeet nests in patches with native trees in Entre Ros was more strongly related to the abundance of preferred nesting sites than foraging sites around the nesting patches (Cavallero 2010). Predation is a factor that has been proposed as influencing selection of nesting sites by monk parakeets because predation of eggs and chicks in nests could be an important cause of mortality at nests ( Navarro e t al. 1992; Sol et al. 1997; Burger and

PAGE 64

64 Gochfeld 2005, 2009 ; Spreyer and Bucher 1998). In this study, all nests were in eucalyptus trees around farm houses and manmade structures inhabited by people, where the abundance of predators of monk parakeet nests is probably low (Burgess and Gochfeld 2009), but human disturbance may have influenced nest density in ways that we did not detected. Presence or absence of human control was not important for explaining density of monk parakeet nests in nesting patches, implying that control measures were not effective in decreasing density of monk parakeet nests. In fact, the relationship between control and density of parakeet nests was positive rather than negative. However, a few farmers in sites with nest density eq ual to zero indicated that they destroyed all parakeet nests and killed or harassed the monk parakeets, and did not let them return again. If control measures had been recorded in a way that measured efficacy, rather than presence/absence of control, this variable may have contributed more to explaining nest density. Management I mplications The number of monk parakeet nests in a farm could be reduced by limiting the available nest sites by removing eucalyptus trees. Potentially, trees with lower height and poor structure to h ol d nests such as native trees of genus Acacia or Prosopis could be used to replace some benefits of eucalyptus (e.g., shade and wind break) without providing attractive nest ing sites for monk parakeets The replacement of introduced eucalyptus trees by native trees also could favor an increase of local biodiversity with additional values for ecosystem services ( Butterfield 1995; Haggara et al. 1998; B urghardt et al. 2008) A lternatively the structure of the eucalyptus trees could be changed so that they do not support nests (Volpe and Arambur 201 1 ). However,

PAGE 65

65 the efficacy of these strategies to decrease density of monk parakeet nests in a farm is unknown. The lack of significance of landscape variables in explaining density of monk parakeet nests in farms with eucalyptus, limits the possibilities of recommending management measures at this level. The only management measurement that emerges from results of this study would be to decrease the amount of tree patches around the farms. H owever, eliminating trees would result in a more simplified landscape, which could have important consequences for biodiversity and ecosystem services provided by trees (Power 2010 ; Batry et al. 2011). Also, the benefits probably would be low, considering the low weight of this variable for explaining density of monk parakeet nests in the farms. Therefore, based on this study, I would not recommend management measures at landscape level for decreasing density of monk parakeet nests in farms with eucalyptus

PAGE 66

66 Figure 31. B uffer extents used for s ampling l andscape level variables from satellite images around each nesting patch (central point). 1 000 m 2000 m 3000m

PAGE 67

67 Table 3 1 Summary of vegetation structure and landscape metrics included in models o f monk parakeet nest density in inhabited farms with eucalyptus in east central Argentina. n=62 nesting patches for local level, n=42 nesting patches for landscape level. Variable Description Mean SE Range Local level EUHGT Mean height of eucalyptus tre es at each site (m) 25.28 0.62 15.32 38.13 EUAREA Area of patch covered by eucalyptus canopy (ha) 0.27 0.04 0.04 1.12 EUPROP Proportion of patch area covered by eucalyptus canopy 0.24 0.02 0.03 0.78 CONTROL Application of control measures on each site t o decrease monk parakeet nesting (presence/absence) NA NA NA Landscape level CRPLAND Percentage of the landscape with preferred food crops (corn and sunflower ) 12.11 0.67 4.14 24.12 PSTPLAND Percentage of the landscape with pastures and other agricult ural uses potentially available to monk parakeets for foraging (including weedy and fallow fields) 17.73 1.09 2.48 32.54 TRPLAND Percentage of the landscape with trees (either native or introduced) 25.18 3.22 0.96 79.59 TRPD Density of patches with trees on the landscape (#/km 2 ) 1.07 0.07 0.28 2.40 These variables were estimated at three buffer extents (1000, 2000 and 3000 m) around each nesting site. Values are given for the 3000 m buffer. Different buffers have different values for metrics, b ut the relationships among variables were qualitatively similar.

PAGE 68

68 Table 3 2 Minimum AIC c mo dels for density of monk parakeet nests in inhabited farms with eucalyptus trees in east central Argentina (n=35) Models were developed with: 1) local variables only, 2) landscape variables at 3 spatial extents, and 3) local and landscape variables in the same models (muti level models). All models with within a level are presented. Models are ordered based on model performance within each level (lower AICc value indicating better model performance). Level Model AICc Akaike weight i ) Local EUPROP EUPROP + CONTROL EUPROP + EUHGT EUPROP + EUAREA EUAREA 16.49 18.30 18.61 18.71 20.47 0.33 0.13 0.11 0.11 0.04 Landscape 1 km 1TRPLAND 1TRPLAND + ( 1TRPD) 1CRPLAND + 1TRPLAND ( 1TRPD) ( 1PSTPLAND) ( 1CRPLAND) 18.15 20.33 20.47 20.72 21.03 21.17 0.36 0.12 0.11 0.10 0.08 0.08 Landscape 2 km 2TRPLAND 2CRPLAND + 2TRPLAND ( 2PSTPLAND) 2TRPLAND + 2TRPD ( 2TRPD) 2CRPLAND 2CRPLAND + 2TRPLAND + 2TRPD 2CRPLAND + ( 2PS TPLAND) 18.91 20.38 20.63 21.20 21.26 21.31 22.83 22.87 0.29 0.14 0.12 0.09 0.09 0.09 0.04 0.04 Landscape 3 km 3TRPLAND ( 3PSTPLAND) 3CRPLAND + 3TRPLAND ( 3TRPD) 3CRPLAND 3TRPLAND + 3TRPD 3CRPLAND + ( 3PSTPLAND) ( 3PSTPLAND) + 3TRPD 3CRPLAND + ( 3TRPD) 3CRPLAND + 3TRPLAND + ( 3TRPD) 19.25 19.72 20.56 21.25 21.26 21.50 21.86 22.04 22.98 23.00 0.23 0.19 0.12 0.09 0.09 0.08 0.06 0.06 0.04 0.04 Multi level EUPROP + 2TRPLAND EUPROP + 3TRPLAND EUPROP + 1TRPLAND 16.62 16.62 16.70 0.34 0.34 0.32 Brackets ind icate a negative relationship with nest density

PAGE 69

69 Table 3 3 Regres sion results for factors considered for predicting density of monk parakeet nests in inhabited farms with eucalyptus trees in east central Argentina. Coefficients and associated standard errors for each predictor variable are derived from multimodel inferences using all parameter subsets and Akaike weights ( i). i for each predictor variable shows the sums of Akaike weights for all possible models in which the predictor variable was inco rporated at each level T he larger the i, the more important a variable is relative to other variables. Variables are ordered within each level based on the importance value ( i). Level Variable Coefficient CI i Local Proportion Eucalyptus area (EUPR OP) 1.58 0.84* 0.72 Human control (CONTROL) 0.04 0.06 0.13 Tree area( EUAREA ) 0.04 0.10 0.11 Tree height ( EUHGT ) 0.002 0.004 0.11 Landscape 1 km Percentage of landscape with tree patches ( 1TRPLAND ) 0.010 0.007* 0.59 Density of tree patches (1 TRPD ) 0.03 0.07 0.22 Percentage of landscape with preferred food crops ( 1CRPLAND ) <0.001 0.005 0.19 Percentage of landscape with pastures (1 PSTPLAND ) 0.001 0.002 0.08 Landscape 2 km Percentage of landscape with tree patches (2 TRPLAND ) 0.008 0.006* 0.57 Percentage of landscape with preferred food crops (2 CRPLAND ) 0.006 0.01 0.31 Density of tree patches (2 TRPD ) 0.008 0.10 0.23 Percentage of landscape with pastures (2 PSTPLAND ) 0.004 0.005 0.17 Landscape 3 km Percentage of landscape with tree patch es ( 3TRPLAND ) 0.006 0.005 0.47 Percentage of landscape with preferred food crops ( 3CRPLAND ) 0.01 0.02 0.34 Percentage of landscape with pastures (3 PSTPLAND ) 0.01 0.01 0.31 Density of tree patches (3 TRPD ) 0.02 0.14 0.29 95% confidence intervals for multimodel weighted coefficients for each of these predictor variables did not include zero, indicating that these factors probably affected abundance of monk parakeet nests.

PAGE 70

70 CHAPTER 4 INFLUENCE OF SOCIO PSYCHOLOGICAL AND SOCIODEMOGRAPHIC FACTORS O N FARMERS PREFERENCES FOR MANAGEMENT STRATEGIES TO DECREASE MONK PARAKEET DAMAGE TO CROPS Background G ranivorous bird species associated with agroecosystems cause damage to crops, feedlots and stored grains worldwide (Pinowski and Kendeigh 1977 ; De Grazio 1978; Feare 1993 ; Bruggers and Zaccagnini 1994). The number of species causing this damage is relatively small, but their impacts often are significant ( Pinowski and Kendeigh 1977; De Grazio 1978; Feare 1993 ; Bruggers et al. 1998). Some species, such as red winged blackbirds ( Agelaius phoeniceus ) in North America, eared doves ( Zenaida auriculata) in South America and red billed quelea ( Quelea quelea) in Africa, comprise flocks and communal roosts of thousands of individuals and perform wide movements (Bel etsky 1996; Bucher 1992a; and Bruggers and Elliot 1989, respectively). Other species that cause damage, including the roseringed ( Psittacula krameri ) and the monk parakeets ( Myopsitta monachus ) although less numerous and more resident, are also very soci al and visible (Spreyer and Bucher 1998; Kahn 2003). The conspicuousness of the se birds and the damage they cause, plus the high variability in damage, make objective estimation of damage by farmers difficult (Conover 2002) and contribute to a tendency to overestimate losses (Dyer and Ward 1977; Bucher 1992a 1998). Consequently, farmers often apply management measures to decrease bird damage that are not economically effective or are contrary to research findings (Bomford and Sinclair 2002; Tracey et al 2 007). The tolerance threshold to bird damage and, consequently, the point at which a farmer decides to apply control measures, are likely to vary depending on psychological

PAGE 71

71 and economic factors (Avery 2002). Although no studies have been conducted compari ng the influence of psychological and economic factors on management decisions about bird pests studies of insect pest management in crops have shown that socio psychological factors, such as perceptions of pest status and attitudes toward control, can in fluence farmers decisions about pest management more than economics, particularly when farmers tend to overestimate crop losses ( Mumford and Norton 1984; Heong and Escalada 1994; Heong and Escalada 1999; Heong et al. 2002) Similarly, studies of other types of humanwildlife conflict have shown that sociopsychological factors influence peoples decisions about management strategies to these conflicts (Pierce et al. 2001) Usually, people with negative attitudes toward wildlife species tend to prefer more invasive control methods to resolve humanwildlife conflicts ( Locker et al. 1999; Coluccy et al. 2001 ; Don Carlos et al. 2009; Loyd and Miller 2010). Also, perceptions of problems with wildlife species, previous knowledge about a management method, and the perception of efficacy of the management method influence preferences for management actions to decrease humanwildlife conflicts (Stout et al. 1993, 1997 ; Locker et al. 1999; Zinn and Andelt 1999; Jonker et al. 2004). In spite of the potential importanc e of socio psychological factors for explaining management decisions to reduce wildlife damage to crops, research in this area is relatively scarce, particularly when related to bird damage (Timm 1991 in Clergeau 1995). Studies evaluating the influence of socio psychological factors on tolerance levels of farmer s to wildlife damage or decisions about management measures to decrease it have focused mostly in conflicts involving mammals, generally deer ( Decker and Brown 1982; Messmer and Shroeder 1996; Campa et al. 1997 ; West and Parkhurst

PAGE 72

72 2002). In addition, no studies have looked at the relationship of socio psychological factors, such as subjective norms (i.e., the perception of social pressure to perform a particular behavior Fishbein & Ajzen 1975; Ajzen 1985, 1991) or perceived behavioral control (i.e., the perception of ability to perform a particular behavior, Ajzen 1985, 1991) on management decisions to decrease wildlife damage to crops. Previous studies have shown a relationship between one or both factors with decisions about insect pest management (Heong and Escalada 1999 ; Heong et al. 2002) and wildlife related activities, such as hunting ( Rossi and Armstrong 1999 ; Hrubes et al. 2001) and the use of devices to decrease humanbear conflicts (Martin and McCurdy 2009) Finally, socio demographic factors that influence preferences for management alternatives to decrease humanwildlife conflicts, such as age and education (Bjerke et al. 1998 ; Koval and Mertig 2004; Loyd and Miller 2010), have not been studied in relation to farmers preferences for management strategies to decrease bird damage to crops. R esearch on factors underlying farmers preferences for alternative management measures to decrease bird damage to crops is needed to develop effective ex tension programs and improve management of bird damage to crops (Bomford and Sinclair 2002 ; Tracey et al. 2007). In this study, I applied a behavioral decision approach to determine farmers preferences for management strategies to decrease damage from monk parakeets ( Myiopsitta monachus ) to crops in Argentina and to evaluate sociopsychological factors and sociodemographic factors that influenced those preferences. In South America, particularly in Argentina, the monk parakeet is considered one of the mos t important bird pest species causing damage to grain crops (Bucher and Bedano 1976; Bucher 1992a,

PAGE 73

73 1992b; Bruggers et al. 1998). Damage involves principally sunflower, corn, and sorghum, with occasional damage to wheat and rice (Bucher y Bedano 1976 ; Bucher 1984, 1992 a ; Bruggers y Zaccagnini 1994) Quantification of damage to crops by monk parakeets is very scarce, but indicates low (< 5%) to moderate (up to 20%) crop loss ( De Grazio y Besser 1975; De Grazio 1985; Canavelli et al. 2008). T raditionally, l ethal control has been applied by government agencies, agricultural professionals and farmers as the most effective method for decreasing monk parakeet damage to crop s in Argentina (Bucher 1984, 1992 a,b) Several methods have been used, including nest bur ning or destruction, shooting, payment of boun ti es trap ping, netting, toxic baits, spraying of nests with insecticides. Since the 1980s, the primary lethal control method has been insecticides mixed with grease and applied on nest openings to produce intoxication and potentially death of birds entering the nest ( Arambur 1991). However, objections to this method are increasing and new methods are required (Canavelli and Zaccagnini 2007; Canavelli and Aramburu, in press ) Additionally, monk parakeets repres ent conflicting values for different groups of people because, although they are considered a pest species, this species also is valued as a domestic pet (Moschione and Banchs 2006). Understanding the bases of farmers preferences for management strategies to decrease monk parakeet damage to crops is crucial for managing conflicts between monk parakeets and crop production (Canavelli and Zaccagnini 2007). Based on the historical context of bird damage management in Argentina, I expected farmers to highly favor population control methods such as lethal and reproductive control, to decrease monk parakeet damage to crops. Additionally, I

PAGE 74

74 expect ed preferences of farmers for specific management strategies to be related to attitudes toward monk parakeets percept ion s of expectations from other people about damage management (subjective norms) and perceived behavioral control about pest management Finally, I also expected other sociopsychological factors, such as perceptions of problems with monk parakeets, know ledge about control strategies and beliefs about the efficacy of management strategies and socio demographic factors, such as age and education, to be related to preferences Methods Study A rea The study was conducted in Paran Department (Entre Ros Prov ince, Argentina), where cattle, milk and crops are the major production activities (Engler and Vicente 2009). Within the department, about 49% of area is devoted to agricultural crops, including in order of importance soybean, wheat, corn, sorghum and sunf lower. This area represents approximately 15% of the total agriculture in the province (Engler and Vicente 2009). Annual mean temperature is 19C and mean rainfall is 1000 mm per year. Paran Department contains 2,314 farms, totaling 488,558 ha (National Agricultural Survey database, INDEC 2002). Most farms (58%) are smaller than 100 ha, following in frequency 100 300 ha (25%), 300500 ha (7%), 5001000 (5%) and > 1000 ha (4%). Farm size is correlated with area devoted to crops within the farm, and commo nly is related to sociodemographic variables, such as land ownership and social organization of work (Engler and Vicente 2009).

PAGE 75

75 Sampling Scheme and Questionnaire D esign I applied a crosssectional study design (de Vaus 2001) based on personal interviews to obtain an instantaneous picture of farmers attitudes perceptions and preferences for alternative management strategies From the target population of farms in Paran department, I selected 115 farmers for faceto face questionnaires. Twenty four farmers were randomly selected among all farmers producing corn and/or sunflower in the 2007 crop season, because these crops are very susceptible to monk parakeet damage (Bucher y Bedano 1976; Bucher 1984, 1992 a ; Bruggers y Zaccagnini 1994). The other 95 far mers were randomly selected from the INDEC 2002 database after stratifying by the area devoted to crops within each farm into four categories: 0.520 has, 2080 ha, 80300 ha, and > 300 ha. To select farmers within each stratum, I ranked farmers based on t he area of the farm devoted to crops. Afterwards, I selected individual farmers using systematic sampling with a random start, in order to cover the size distribution of crop area within each stratum. The number of farmers selected in each stratum was proportional to the total number of farmers in the stratum The questionnaire for the interview was organized in five main sections : 1) farmer s perceptions of bird abundance and damage; 2) farmers knowledge and attitudes toward monk parakeets ; 3) farmers knowledge and preferences for management strategies to decrease or prevent monk parakeet damage to crops 4) personal and external influences on the application of management strategies and 5) socio demographic information (Appendixes B and C ). Most questions in each section were closed and/or structured questions that were completed with the help of the farmer. Opinion responses were categorized with a symmetric scale with a central neutral

PAGE 76

76 category (agree, neutral, disagree) and options for undecided and not answering ( Reiter et al. 1999; Jacobson et al. 2003). Prior to implementation, questionnaires were reviewed by extension agents and agronomists at the Paran Experimental Station at the National Institute of Agricultural Technology (INTA) to determine t he clarity of questions, completion time, and other aspects of survey completion (Reiter et al. 1999). The questionnaire also was pre tested with a sample of seven farmers closely related to the station and adjusted subsequently A summary of the questionn aire structure and content is included i n Appendix B and the final questionnaire in Spanish is included in Appendix C Variable M easurement Preference for m anagement s trategies For evaluating farmers preferences for management strategies to decrease monk parakeet damage to crops I employed the method of paired comparison s, a widely used method to evaluate preferenc e dimensions in psychology, marketing, policy and economics, based on offering the respondent items from a choice set in pairs (Burgess et al. 2002; Brown and Peterson 2003) For each pair of options, the respondent is requested to choose the superior item Each choice is assumed to be independent of all other choices (Brown and Peterson 2003). I evaluated preference of farmers for each of seve n management strategies for decreasing monk parakeet damage to crops: 1) lethal control by shooting, trapping, or poisoning, 2) crop protection with physical or chemical deterrents, 3) agricultural practices, such as early planting, high crop density, etc. 4) habitat management by modifying tree structure, using decoy plots, etc., 5) reproductive control by removing or burning nests, egg oiling, etc. 6) capture of birds and relocation, and 7) integrated pest

PAGE 77

77 management (IPM), defined for the purpose of this study as the integration of several control tactics to reduce the status of the pest (Kogan 1998). Based on these seven strategies, I had a set of 21 pairs of strategies, and respondents were forced to choose one of the two options in each pair. As a sh ort term measure of reliability of the answers, I had two repeated pairs of choices with reverse order of items. These two pairs were randomly selected from the full set of pairs. I randomly sorted the 23 pairs, made ten versions of the questionnaire that differed only in the order of the presentation of pairs, and randomly choose the version to use in a given interview. Altering the order of presentation of pairs has been suggested as a way of decreasing bias on stated preferences due to the order in which pairs are presented and consequently, increasing reliability in the answers ( Brown and Peterson 2003). Sociopsychological factors Perceptions of problems with monk parakeets. Perceptions of problems with monk parakeets included questions about monk pa rakeet damage, population abundance, and trends in damage and abundance in the three years prior to the interview (Appendixes B and C ). Because measures of perceptions of monk parakeet abundance were significantly related with perceptions of monk parakeet damage (Fisher Exact Test p=0.025), and perception of population trends of parakeets were significantly related with perceptions of trends in damage (Fisher Exact Test p< 0.0001), I focused the analysis on perception of damage. Also, I asked farmers about their tolerance to monk parakeet damage on crops. For each crop, I asked if the farmer usually produced the crop in the farm and, if so, whether parakeet damage usually occurred on the crop. In the case of a positive answer about damage, I asked if he/she considered damage as toler able (2) or intolerable (3). For statistical analyses, I combined answers for the three

PAGE 78

78 crops most attractive to monk parakeets (corn, sunflower and sorghum) for each farmer, p arakeets. T o obtain an index of tolerance of damage for each farmer, I averaged tolerance levels among the crops for which farmers reported damage. Index of tolerance varied between 2 (tolerable) and 3 (intolerable). Finally, I asked farmers about their perception of the importance of losses by monk parakeets compared with losses caused by other factors, such as climate, insects, weed, diseases and harvesting machinery. Importance of monk parakeet losses compared to other crop loss causes was captured in th ree categories (1=less important, 2=equally important, 3=more important). In order to have a single value of relative importance of monk parakeet damage compared to other crop losses for each farmer for statistical analyses, I added the importance value of monk parakeet damage compared to all other factors (climate, insects, weed, diseases, harvesting machinery) for each farmer (index range: 5monk parakeet damage considered lower in importance to 15monk parakeet damage considered higher in importance). K nowledge about management strategies and beliefs about their effectiveness. All farmers interviewed knew about monk parakeets and the damage they cause on food crops. Previous knowledge of farmers about management strategies was evaluated in general and for each of seven management strategies: lethal control, crop protection, agricultural practices, habitat management, reproductive control, capture of birds and relocation, and integrated pest management On each case, knowledge was recorded as a binary vari able (1=yes, 2=no). Based on the strategies previously known by each farmer, I also determined the perceived effectiveness for each strategy separately as

PAGE 79

79 well as which strategy farmers considered the most effective. A first question asked about the percei ved effectiveness of each previously known strategy (1=no effective, 2=slightly effective, 3=very effective), and a second question asked farmers to rank the management strategies based on their effectiveness (first, second and third most effective strateg ies). Attitudes toward monk parakeets and management of monk parakeet damage to crops. Attitudes toward monk parakeets were measured using a Likert scale index based on 14 belief items that were successful in separating opinions from different groups of f armers. A Likert scale is an instrument widely used to measure levels of theoretical constructs, such as opinions, attitudes or beliefs, not readily observed by direct means (DeVellis 2003). It includes a set of items that are combined into a composite sco re (De Vellis 2003) Originally, I evaluated a pool of 40 items developed following DeVellis guidelines (2003, pgs. 60101) for assessing attitudes toward monk parakeets and another pool for assessing attitudes toward management of monk parakeet damage to crops. However, after evaluation of item performance with a sample of 12 farmers, I developed a 14item Likert scale only for evaluating attitudes toward monk parakeets, because items for evaluating beliefs about management of parakeet damage to crops wer e not good at separating opinions from different groups of farmers (conservationists vs. productionists). Attitudes ranged from 14, indicating farmers had a less favorable or negative attitude toward monk parakeets, to 42 indicating farmers had a more f avorable or positive attitude toward monk parakeets. Subjective norms about monk parakeet damage management. To determine subjective norms I asked farmers to state what they thought specific reference groups

PAGE 80

80 expected the farmer to do about monk parakeet control (norm belief) and how much farmers cared about expectations of each reference group (motivation to comply, Heong and Escalada 1999). The six reference groups were neighbors, spouse, extension agents from 1) a cooperative, 2) the national government and 3) the state government, and sales agents from chemical companies. Norm belief was scored for each reference group asking the farmer to state if each group expected the farmer to never control monk parakeets (score=1), occasionally (once e very 2 year s, score=2), frequently (at least once a year, score=3) or very frequently (every season, score=4). Motivation to comply was scored for each reference group asking the farmer to state how much he/she cares about the opinion of each group (1= does not care, 2= cares moderately, 3=cares a great deal ). For statistical analyses, a composite measure of subjective norm was estimated for each farmer as the product of norm belief and motivation to comply for each reference group, added for all groups (Heong and Escal ada 1999; Heong et al. 2002). When a reference group did not apply for a farmer (e.g., spouse) or the farmer was not decided about the answer for a specific reference group, a score equal to 0 was used so the group did not counted in the overall sum Compo site scores of subjective norms ranged f rom 1 indicating farmers experience no social pressure to control monk parakeets, to 61, indicating farmers experienced strong social pressure to control monk parakeets. Perceived behavioral control about bird pest management. Perceived behavioral control about bird pest management was evaluated as the perceived confidence of farmers in their own abilities to apply management strategies to decrease bird damage to crops (internal factors) and the influence of external factors limiting these abilities

PAGE 81

81 (e.g., limited access to control devices, high complexity of the current techniques, elevated cost of actual techniques, etc., Appendix C ). Confidence was captured in three categories (1=insecure, 2=moderately secure, 3=very secure) and a composite measure of confidence for each farmer was estimated as the sum of all values for all abilities for each farmer. Similarly, the influence of external factors was also measured in three categories (1=not limiting, 2=moderately limi ting, and 3= highly limiting). The overall value of perception of limiting factors for each farmer was estimated as the sum of the individual values for each item. Finally, f or statistical analyses, a composite measure of perceived behavioral control was e stimated as the product of perceived confidence and limiting factors. Sociodemographic factors S ocio demographic factors also were evaluated for each farmer including age, educational level, area of the farm area of farm devoted to crops and social par ticipation. Social participation was evaluated as the affiliation with one or more farmers organization (0=no affiliation, 1= 1 organization, 2= 23 organizations, 3= 4 5 organizations, 4= > 5 organizations) and the degree of participation to farmers meeti ngs (0=no participation, 1=less than one every 2 months, 2= monthly or bi monthly, 3= > once a month) In order to have a single value for each farmer representing social participation, a composite index was built by adding the affiliation with farmers or ganizations and the degree of participation on farmers meeting (range:07) Other socio demographic factors, such as property ownership, income from crops, and sources of information about pest management strategies were included on the questionnaire for g eneral description of the population but were not used for relating to preferences for management strategies because of incomplete information for all

PAGE 82

82 farmers (property ownership and income from crops) or unclear expected relationship with preferences (sources of information). Statistical A nalyses Preference scores and reliability With data from all respondents, I estimated scale values of preference for each management strategy, developed a scale of preferences, and estimated individual respondent reliabil ity for the answers. I constructed preference matrices for each farmer with the full set of choices (i.e., 7 management strategies, 21 pairs). Each matrix had seven rows and seven columns, one for each management strategy, and each cell represented the pre ference of one strategy compared to another (0= not preferred, 1=preferred). Based on the preference matri x for each farmer I estimated a preference score for each strategy, representing the number of times a strategy was preferred to all other strategies in the set of choices (Brown and Peterson 2003). Finally, I estimated a scale value of preference for each management strategy and its associated standard error by applying the Bradley Terry (BT) model for paircomparison data to the aggregated preferenc e matrix for the sample of farmers. The BT model was run in R with the BradleyTerry2 addon package ( version 0.92 2010, Turner and Firth 2011). As a measure of reliability of the responses, I calculated the coefficient of consistency for each respondent using the number of circular triads (i.e., the number of intransitive responses) for each individual ( Burgess et al. 2002; Brown and Peterson 2003). The coefficient of consistency varies between 1 (no circular triads) and 0 (maximum possible number of triads, Brown and Peterson 2003).

PAGE 83

83 Assessment of the relationship between preferences for management strategies and socio psychological and sociodemographic factors Preference scores for each management strategy were re categorized as : 1=low preference (pref erence scores 0, 1 and 2); 2=medium preference (preference scores 3 and 4); and 3=high preference (preferences scores 5 and 6). Also, because of the large number of categories for many variables relative to the sample size (n=111) I re categorized indepen dent variables with more than four levels and/or uneven distribution of cases amon g categories to variables with three or four levels using probabilistic methods based on frequency distribution (e.g., quartiles) and/or social based criteria (e.g., grouping of age < 40, 4049, 50or educationincomplete primary school, primary school, secondary school and university instruction levels). As all variables were categorical, I compared each sociopsychological and sociodemographic factor wit h preferences for each management strategy using bivariate Chi square analyses. For cell counts of expected values below five I used Fisher's exact test (Agresti 2002) to evaluate statistical significance Additionally, I used proportional odds logistic r egression models in SAS (v.8, 2006) to determine which socio psychological and sociodemographic factors were most strongly related with preferences for each management strategy (Agresti 2002) The relative importance of each variable for predicting prefer ences for each management strategy was evaluated by considering the Akaikes Information Criterion (AIC, Burnham and Anderson 2002) by ranking the variables in order of lowest AIC value (the best performing independent variable) to the highest (the worst performing independent variable, Hagy et al. 2008). Models with of were considered competitive models (Burham and Anderson 2002). Additionally, I used percent concordance, a measure of association of

PAGE 84

84 predicted probabilities and observed responses, as a complementary measure of model fit. I modeled each management strategy separately because evaluation of all strategies simultaneously was too complex. Finally, I explored the correlation between behavioral factors (attitudes, subjective norms and perceived control) and other independent variables using Ch i square tests. For graphical representation of contingency tables, I used mosaic displays when two or three variables were involved (Friendly 2000). When more than three variables were related simultaneously, I used multiple correspondence analyses (MCA) to graphically represent the relationships. Results Preferences for B ird Pest Management S trategies Strategies for monk parakeet population control, such as reproductive and lethal control, were most preferred by farmers, followed by integrated pest manage ment, crop protection, agricultural practices, and habitat management ( Figure 4 1). P references for different strategies were related. For instance, w hen farmers highly preferred reproductive control, they also preferred lethal control and had low preferences for other alternatives, such as crop protection, integrated pest management or agricultural practices ( group A in Figure 4 2 ) Conversely, when alternative strategies to population control were preferred, such as crop protection or agricultural pract ices, farmers had low preferences for all population control strategies (lethal and reproductive control, group B in Figure 4 2 ). Reliability of preference scale values. Farmers averaged 1.26 ( 0 16 SE ) circular triads of a possible maximum of 14. Average coefficient of consistency for preferences was 0.91 ( 0.01 SE). Forty seven percent of farmers (n=52) had no circular triads (i.e.,

PAGE 85

85 inconsistencies on his/her choices), 40% of farmers had three or less triads, and 13% had four or more. Factors R elated t o P references for M anagement S trategies Sociopsychological factors Perceptions of problems with monk parakeets. Most farmers (68%) had experienced damage by monk parakeets on their crops on the last three years. For these farmers (n=75), damage was gener ally insignificant (32% of farmers) or moderate (49% of farmers). Only 19% of these farmers (equivalent to 13% of all farmers) considered damage intense. Damage for the 20062007 crop season was reported as less than 5% crop loss by 32 farmers (29% of all farmers), between 5 and 10% crop loss by 26 farmers (23%), between 10 and 25% by 9 farmers (8%) and greater than 25% by 8 farmers (7% of all farmers). Most farmers reported an increasing trend in damage, with damage equal or greater in the last season comp ared to the average annual damage over the last three years (43% and 41% of farmers, respectively). Most farmers (67%) growing attractive crops for monk parakeets (corn, sunflower or sorghum) experienced damage by monk parakeets on at least one of the crops. However, when damage was observed, most farmers considered the damage as tolerable (76%). Farmers growing other crops (e.g., soybean, alfalfa, millet and flax) mostly reported no damage by monk parakeets, with exception of wheat, where half of farmers growing this crop (n=33) reported some damage by monk parakeets. In this case, most farmers that reported damage by monk parakeets (88%) also considered damage as tolerable. Most farmers (> 60%) considered crop losses caused by monk parakeet damage to be m ore important than losses by weeds, diseases, and harvesting machinery. In

PAGE 86

86 contrast, most farmers (> 60%) considered crop losses caused by adverse climate and pest insects equally or more important than damage by monk parakeets. During the interviews, some farmers mentioned losses by weeds and diseases also could be more important than damage by monk parakeets, but these two problems could be managed well, so that losses were lower. The mean and m edian score of perception of importance of monk parakeet were identical ( 11; range = 5 15) suggesting farmers perceived losses by monk parakeets to be relatively important compared to other causes of crop loss. Preferences of farmers for management strategies generally w ere not associated with perceptions of damage by monk parakeets to crops (Table 4 1). The only statistically significant relationship was that farmers who perceived monk parakeet damage as more important than other crop losses highly preferred lethal control as a management strategy to decrease damage (Table 4 1). N o variable representing perception of problems with monk parakeets was included among the most important variables explaining preferences for any management alternative (Table 4 2). Knowledge about management strategies and beliefs about the ir effectiveness. Most farmers (88%) knew some control strategies for decreasing monk parakeet damage before they were intervi ewed, some of them did not (12% ). Lethal control, reproductive control and agricultural practices were commonly known by farmers ( Figure 4 3 ). Integrated pest management (IPM) and capture and relocation of birds were the l east known alternatives by farmers ( Figure 4 3 ). Lethal and reproductive controls also were considered the most effective strategies by most farmers, while agricul tural

PAGE 87

87 practices and integrated pest management followed those in order of importance ( Figure 4 4). Preferences of farmers for management strategies were clearly associated with previous knowledge about the strategy and its perceived efficacy, with farmers generally preferring strategies known to them or perceived as effective before the interview. For instance, farmers who knew about reproductive control, lethal control, integrated pest management, and capture and relocation before the interview highly pre ferred these management strategies (Table 4 1). Similarly, farmers who perceived these strategies, as well as crop protection and agricultural practices, very effective also preferred them (Table 4 1). However, these relationships were not observed for habitat management. Both previous knowledge and perceived efficacy of the method were important variables explaining preferences for integrated pest management (Table 4 2). Previous knowledge also was one of the two most important variables explaining prefere nces for capture and relocation (Table 4 2). Previous knowledge about a strategy was a prerequisite for farmers to respond to questions about perceived efficacy. Therefore, these variables were related. Attitudes toward monk parakeets Attitude scores indicated that farmers predominantly feel a negative attitude toward monk parakeets ( mean score=24, m edian score=22, range=1442) Most of farmers (> 60%) disagreed with all positive statements about monk parakeets. In contrast, opinions were more closely s plit between agreed and disagree for negative statements about parakeets, with the exception of two statements. Most farmers agreed with the following statements: Monk parakeets only

PAGE 88

88 make me lose money (65% of farmers) and Monk parakeets bothered me bec ause they decrease my crop production (83% of farmers). Attitude was the most important factor explaining monk parakeet preferences for management strategies to decrease monk parakeet damage to crops for four of the seven strategies evaluated: reproducti ve control, lethal control, crop protection and agricultural practices (Table 4 2). Farmers with negative attitudes toward monk parakeets highly preferred population control strategies (reproductive and lethal control, Table 4 1 ). In contrast, farmers with positive attitudes toward monk parakeets highly preferred alternative management strategies, such as crop protection and agricultural practices (Table 4 1). Also, attitude was one of the two most important factors for preferences for capture and relocation (Table 4 2). Attitude was not important for explaining preferences for integrated pest management (Table 4 2). Subjective norms about monk parakeet control Farmers recognized external influences from different groups of people (spouse, neighbors, exten sion agents from cooperatives, government and agrochemical companies) on what it is expected for them to do about monk parakeet control. However, the influence of these groups usually was considered unimportant in their decision about control. More than 60 % of the farmers thought that all groups would expect them to control monk parakeets frequently (at least once a year) or very frequently (once every season). However, the opinion of these groups of people did not concern farmers very much, especially if c oming from neighbors or commercial agents from agrochemical companies. Usually, between 40 and 50% of farmers did not care about the opinion of each reference group, and more than 60% of farmers did not care about the opinion of neighbors or commercial age nts

PAGE 89

89 from agrochemical companies. F armers experienced only moderate social pressure regarding the management of monk parakeet problems (mean score of subjective norms =29.33, median=28, range=161). Similar to attitudes, farmers who perceived strong social pressure to control monk parakeets highly preferred population control strategies (reproductive and lethal control) as well as habitat management (Table 4 1). No statistically significant relationships were found with preferences for other management strategies (Table 4 1) Also, subjective norms were nonincluded among the most important factors explaining preferences for any management strategy (Table 4 2). Perceived behavioral control about bird pest management M ost farmers (88%) were confident in identi fying pest bird species, but fewer ( < 6 0%) were confident about knowing multiple alternatives for bird pest management, the biology of these birds, existing regulation o f bird pest management, or methods for capturing bird pests, among other abilities. All external factors presented to farmers as limiting their abilities for applying bird pest management were recognized by most of farmers (>60%) as moderately or highly limiting, including limited access to information and/or limited available information, hi gh complexity of current techniques, and difficulties with cost benefit evaluations. Farmers answers were divided only when we asked about community opinion regarding certain management strategies (e.g., lethal control), because half of farmers did not consider community opinion as a limiting factor for the application of those strategies. When farmers confidence and the influence of limiting factors were integrated, median score of the composite measure of perceived behavioral control was 153 and mean sc ore was 162.45 (range: 42336), suggesting that farmers

PAGE 90

90 feel moderate control i n applying management strategies to decrease bird damage, probably as a consequences of relatively low moderate internal control and high limitation of external factors. The pe rception of behavioral control was related only to preferences for integrated pest management (Table 4 1). Farmers who perceived high control i n applying management strategies to decrease bird damage highly preferred integrated pest management, an alternat ive and more complex strategy to population control. Similar to subjective norms, perceived behavioral control was not included among the most important factors explaining preferences for any management strategy (Table 4 2). Sociodemographic factors Most of interviewed farmers (n=111) were male (97%). All of them were 18 years old and older, most of them between 30 and 59 yrs old (71% of total). Most farmers had finished primary school (54%) but fewer finished secondary school (16%) or university ( 17%). Th e area under management by each farmer varied between 16 and 3200 ha, with most of farmers (6 0%) making decisions for 300 ha or less. About half of farmers (n=49) conducted production activities on their own property and a similar number (n=44) rented land for production. Only 49% of the farmers had greater than 25% of the farm area in crops. However, income from crops was important (50% or more of the total income for the farm) for many farmers (n=33) who answered the question (n=87). Most farmers (82% ) conducted production activities on the farms besides crops (cows, milk production, etc.). Most farmers (60%) did not belong to any farmer organization. The ones who did (n=44), mostly were associated with one organization (59%). None of the interviewed

PAGE 91

91 farm ers belonged to a conservation organization. Most farmers (80%) participated in farmers meetings and conferences, usually once every two months or less. Preferences for management strategies to decrease monk parakeet damage were generally less related to socio demographic factors than sociopsychological factors. However, young farmers ( < 40 yrs old) with high levels of formal education (university instruction) had lower preference for population control methods (reproductive and lethal control) compared to older farmers with less formal education, and the reverse occurred for agricultural practices (Table 4 1). Farm area, percentage of area devoted to crops, and social participation were each related to preferences for reproductive control. F armers with s mall farms, a smaller amount of area devoted to crops, and elevated social participation, highly preferr ed this management strategy (Table 4 1). Also, small farmers had low to moderate preference for habitat management (Table 4 1). N o socio demographic fac tor w as included among the most important variables explaining preferences for any management alternative (Table 4 2). Relationships among S ociopsychological and S ociodemographic F actors Attitudes were related with socio demographic factors, such as age and education (Table 4 3). Most farmers with negative attitudes toward monk parakeets were 60 yrs old and relatively less educated farmers (incomplete primary school, Figure D 1, Appendix D ). Similarly, attitudes were related to sociopsychological factors, specifically beliefs about the effectiveness of management strategies M ost farmers wi th negative attitudes toward monk parakeets consider ed lethal and reproductive control as the most effective strategies ( Figure D 2, Appendix D ). Beliefs about the effectiveness of management strategies also were related with subjective norms about monk parakeet control, with most farmers considering population control strategies (lethal and

PAGE 92

92 reproductive control) as the most effective strategies perceiving moderate to high social pressure ( Figure D 3, Appendix D ). Perceived control regarding bird pest manag ement was not related with any other sociodemographic or sociopsychological factor (Table 4 3). Results S ummary In summary, socio psychological factors, such as attitudes toward monk parakeets, previous knowledge of each strategy and beliefs about effectiveness of each strategy, were strong ly related to preferences for management strategies (Tables 41 and 42) Other sociopsychological factors, such as subjective norms about monk parakeet control and perceived control about bird pest management were r elated to preferences for management strategies to a much lesser extent. Finally, perceptions of problems with monk parakeets were not related to preferences for management strategies, with the exception of the perception of i mportanc e of monk parakeet dam age compared to other loss causes, which was strongly related to preferences for lethal control (Table 41) Socio demographic factors were less related to preferences for management strategies than were socio psychological factors. Only age and level of f ormal education were related to preferences for population control methods (reproductive and lethal) and agricultural practices (Table 41) Other socio demographic factors (farm area, percentage area devoted to crops and social participation) were related only to preferences for reproductive control (Table 41) The most preferred strategies, reproductive and lethal control, were related to the greatest amount of factors evaluated in this study (n= 9 and 7, respectively Table 41 ). However, some of these factors were related among each other ( Table 43) Attitudes

PAGE 93

93 were related with socio demographic factors, such as age and education and to socio psychological factors, specifically beliefs about the effectiveness of management strategies ( Table 43) Also b eliefs about the effectiveness of management strategies were related with subjective norms about monk parakeet control (Table 43). Discussion P references of Farmers for Management Strategies and Factors Related with those P references Population control strategies, such as nest destruction and killing of birds, were perceived by farmers as the most effective strategies for decreasing monk parakeet damage to crops and also were the most preferred strategies. The use of these strategies is historical in Ar gentina (Bucher and Bedano 1976; Bucher 1984; Arambur 1991; Bucher 1992 a and b ; Bruggers et al. 1998), including in the region of this study (Zaccagnini and Bucher 1983 ; Gimnez and Salomn 2000). Although the effectiveness of these strategies to decreas e damage to crops by monk parakeets has not been evaluated, intense campaigns of population control, where monk parakeets are killed on the nests with pesticide can produce considerable reductions in monk parakeet populations (Bucher 1985, 1992b). The redu ction in populations produced by lethal control or nest destruction, together with a perception of a positive relationship between the application of these control measures and a decrease on monk parakeet damage to crops (Bucher and Bedano 1976; Zaccagnini and Bucher 1983), could explain the perceptions of effectiveness and preference for both methods by farmers. However, other studies suggest that neither of these strategies alone would be cost effective for decreasing monk parakeet damage to crops because of the difficulties of producing a large enough decrease in the population (Canavelli 2003) with a cost lower than the

PAGE 94

94 damage (Bucher 1992b). Also, these methods are not environmentally safe, either for monk parakeets or other species (Keith 1991; Bucher 1992b ; Zaccagnini 2006). Therefore, the use of lethal or reproductive control as the main strategies for decreasing monk parakeet damage to crops is at least questionable, not only by wildlife biologists but also by some farmers and the general public, who have shown strong opposition to population control of monk parakeets in some cases (Canavelli and Arambur, in press ). Preferences of farmers for management strategies were more strongly related to attitudes toward monk parakeets than to any other socio p sychological or sociodemographic factor. Similar to what has been found in previous studies looking at the influence of attitudes on preferences for management actions involving wildlife species (e.g., Bjerke et al. 1998 ; Stout et al. 1997 ; Don Carlos et al. 2009; Loyd and Miller 2010), m ost farmers with negative attitudes toward monk parakeet preferred invasive population control methods, such as lethal and reproductive control, and farmers with positive attitudes toward monk parakeets preferred nonletha l strategies, such as crop protection and agricultural practices Given crop damage from monk parakeets was considered tolerable in this study the predominantly negative attitudes toward monk parakeets may be related to past problems more than the percept ion of actual damage (Zinn and Andelt 1999) This p articularly would be the case if those problems contributed to building strongly held attitudes that are relatively stable and difficult to change (Pierce et al. 2001). This proposition may be supported by the relationship of attitudes toward monk parakeets with age, with older farmers having predominantly negative attitudes compared to y ounger farmers

PAGE 95

95 Preferences of farmers for management strategies also were strongly associated with perceived efficacy o f management strategies and, to a l esser extent, to previous knowledge about those strategies. Farmers who were familiar with a management alternative to decrease monk parakeet damage to crops, and perceived the management strategy to be effective also pre ferred this strategy. However, some disparities were observed in the relationships between preferences and knowledge or perceived efficacy. For example, although i ntegrated p est m anagement was the least known management strategy (< 20% of farmers knew about it) this strategy ranked third in preference, following reproductive and lethal control strategies i n the general scale of preferences/ranking. Similarly, many farmers preferred integrated pest management although this strategy was not perceived to be a mong the most effective ones. In addition, some farmers perceived strategies such as agricultural practices as very effective, but they were not highly preferred. Consequently although knowledge and perception of efficacy were important predictors of pref erences for some management strategies, they were less directly related t o preferences than attitudes Other socio psychological factors, such as subjective norms and perceived control, were related to preferences of farmers for management strategies but t o a less er degree than attitudes perceived efficacy of management strategies or previous knowledge about those strategies Subjective norms and perceived control also have been found to predict human behavior regarding other types of humanwildlife intera ctions (e.g., hunting) although usually to a lesser degree than attitudes (Rossi and Armstrong 1999 ; Hrubes et al. 2001; Martin and McCurdy 2009). Finally, socio demographic variables, such as age and education, also had little influence on preferences of farmers for

PAGE 96

96 management strategies. However, some disparities occurred depending on the management strategy. For example, age and education were significantly related with preferences for population control methods, particularly lethal control. Similarly, other studies have found a relationship between supporting the extirpation of wolves and age and educational level (Bjerke et al. 1998), or educational level, but not age, and preference for lethal control of feral cats (Loyd and Miller 2010). Given attitudes toward wildlife generally are related to educational level and age (Kellert 1980; Kellert and Berry 1987 in Loyd and Miller 2010), the relationship between preferences for management strategies to decrease monk parakeet damage to crops and age and educ ation is probably mediated through a direct relationship of age and education with attitudes and an indirect relationship with preference for a strategy, as has been proposed for behavioral intentions in general (Fishbein and Ajzen 1975; Ajzen 1991). Contr ary to what I expected, preferences for management strategies, including lethal or reproductive control, were not related with perceptions of magnitude of damage by monk parakeets. Previous studies have found a direct relationship between perception of dam age and decisions in pest management (e.g., Savary 1993; Heong and Escalada 1999, 2002) and perceptions of risks and preference for wildlife management techniques (e.g., Stout et al. 1997 ; Coluccy et al. 2001). In this study, t he proportion of interviewed farmers reporting damage by monk parakeets was substantial (68%) indicating wide spread consumption of crops by monk parakeets in the region However, the reported magnitudes of damages were relatively low ( < 10% of crop loss), in agreement with previous studies of damage perception (Zaccagnini and Bucher 1983; Gimnez and Salomon 1999, 2000) and quantitative evaluations of damage in the

PAGE 97

97 r egion (Zaccagnini y Cassani 1985; Zaccagnini y Tate 1991, 1992; Gimnez and Salomn 1999; Canavelli et al. 2008). Addi tionally, most farmers were tolerant of monk parakeet damage, even in very susceptible crops, such as corn, sunflower or sorghum. T he perception of low magnitude of damage by monk parakeets to crops and this tolerance to damage may explain the lack of significant relationships between these factors and preferences for management alternatives. Nevertheless, some farmers still perceived monk parakeet damage as an important cause of crop loss, and even more important than other sources. Beliefs about the impor tance of damage by monk parakeets compared to other crop losses was the only factor that was significantly related with preferences for lethal control, suggesting that when damage from parakeets is considered greater than other causes of crop loss, farmers prefer a management strategy that produces a decrease in the bird population. Previous studies also have found that people are more willing to accept more invasive population control methods, particularly lethal control, when the severity of incidents wit h wildlife increases (e.g., Bjerke 1998; Zinn et al. 1998; Locker et al. 1999 ; Don Carlos et al. 2009). Management I mplications T he monk parakeet is considered one of the most important bird pest species causing damage to grain crops in Argentina (Bucher and Bedano 1976 ; Bucher 1992a, 1992b; Bruggers et al. 1998). Currently, the main management strategies are population control methods such as lethal control with insecticides mixed with grease and applied on nest openings or capturing parakeets and destroyi ng the nests. In this study, farmers preferred reproductive control more than lethal control as a management strategy to decrease monk parakeet damage to crops However, it is possible that reproductive control masked some ways of lethal control in the nes ts, such as burning nests with

PAGE 98

98 nestlings inside. O bjections to th e se method s are increasing by some farmers and the general public, who have shown strong opposition to lethal control of monk parakeets in some cases ( Canavelli and Aramburu, in press ). Consequently new methods are required ( Canavelli and Zaccagnini 2007; Canavelli and Aramburu, in press ). Se veral management strategies other than lethal or reproductive control are currently available to prevent monk parakeet damage to crop fields or protect t hose fields from damage, including agricultural practices, such as increasing crop density and sowing deterrent crops, or using bio repellents (Canavelli and Arambur, in press ). Unfortunately, no evaluations have been made o f the efficacy of these managem ent strategies to decrease monk parakeet damage to crops. In order to shape farmers opi nions and management decisions about strategies other than lethal or reproductive control, evaluations of the efficacy of these strategies are needed in the regions where conflicts with monk parakeets are important. Given current uncertainties in the outcome of management actions to decrease monk parakeet damage to crops, it would be useful to adopt an adaptive management approach ( Holling 1978 ; Walters 1997 ; Shea et al 2002; Parkes et al. 2006) in which multiple land owners are involved and experiments with different management options are conducted in the region (Canavelli and Zaccagnini 2007; Canavelli and Arambur, in press ). Involving stakeholders, farmers in this case, in management actions and management decisions in the field would increase ownership of stakeholders in the results and enhance credibility of management agencies coordinating the activities ( Messmer et al. 1997). Additionally, f ield projects demonst rating effectiveness of different practices to decrease monk parakeet damage to crops likely would stimulate

PAGE 99

99 opinion changes about management and/or encourage adoption of new practices (Stout et al. 1997; Tracey et al. 2007). E xtension actions within a management program would have to integrate people with different points of view, in order to avoid public controversies about management, particularly in regional management programs and/or programs based on controversial management strategies, such as lethal or reproductive control. In this study, attitude was the main driver of preferences for management strategies. Additionally, a diversity of attitudes toward monk parakeets w as observed among farmers Although farmers with a negative attitude predominated, about 45% of the farmers had a moderate or positive attitude toward monk parakeets. The observed diversity of attitudes toward monk parakeets is similar to attitudes of farmers tow ard other wildlife species (wolves deer, prairie dogs, etc) that damage properties (Bjerke et al. 1998 ; Loker et al. 1999 ; Zinn and Andelt 1999; Jonker et al. 2004). A ttitudes could be based on ethical issues, involving the prioritization of different values (Stout et al. 1997) as well as previous experiences and/or worries abou t possible damage in the future instead of the real experiences ( Brown et al. 1978) Results from this study reinforce the importance of understanding the underlying factors supporting farmers attitudes as well as addressing heterogeneity in farmers attitudes toward monk parakeets in extension activities oriented to increase farmers preferences for management strategies other than population control. Additionally, extension activities would have to focus on changing attitudes, as a prerequisite of changing behavior, more than increasing knowledge about management practices, determining and

PAGE 100

100 communicating levels of damage or showing the efficacy of alternative management strategies. Finally, research and e xtension actions also need would need to be proactive, so that intolerable losses are anticipated and avoided instead of trying to eliminate a situation once it has occurred (Fritzell et al. 1997). In this study, an increasing preference for lethal control was observed as perception of damage intensity increased compared to other crop losses. This relationship may indicate that, if problems increase in the future and farmers tolerance of damage decreases, farmers may be more likely to support lethal methods to resolve these problems (Fritzell et al. 199 7 ; Loker et al. 1999). Therefore, it would be worthwhile to conduct research and extension activities focused on monitoring the intensity of damage and abundance of monk parakeets and communicating the results to the general public. Although these activiti es probably would not influence farmers preferences for management strategies, at least with the current perception of low levels of damage, they would be useful in anticipating intolerable levels of damage and reducing application of lethal strategies to decrease this damage.

PAGE 101

101 Management strategy rank 1 2 3 4 5 6 7 Scale values 1 2 3 4 5 6 7 RC LC IPM CP AP HM CR Figure 4 1 Ranking of preference for management alternatives ( s.e.) by farmers (n=111). Scale values on the graph were estimated with the Bradley Terry model. RC= reproductive control LC= L ethal control, IPM= Integrated pest management, CP= Crop protection, AP= Agricultural practices, HM= Habitat management, CR= Capture and relocation.

PAGE 102

102 LC CP AP HM RC CR IPM -0.98 -0.26 0.46 1.17 1.89 Axis 1 -0.98 -0.26 0.46 1.17 1.89 Axis 2 L M H L M H L M H L M H L M H L M H L M H L M H L M H L M H L M H L M H L M H L M H LC CP AP HM RC CR IPM A B LC CP AP HM RC CR IPM -0.98 -0.26 0.46 1.17 1.89 Axis 1 -0.98 -0.26 0.46 1.17 1.89 Axis 2 L M H L M H L M H L M H L M H L M H L M H L M H L M H L M H L M H L M H L M H L M H LC CP AP HM RC CR IPM A B Figure 4 2 Distribution of the levels of preferences for each management strategy based on a multiple correspondence analysis (MCA). LC= lethal control, CP= crop protection, AP= agricultural practices, HM= habitat management, RC= reproductive control, CR= capture and release, IPM= integrated pest management. Levels of preference: L= low, M= medium, H= high. Group A represents high level of preference for population control methods (lethal and reproductive control) and low preferentes for crop protection and agricultural practices. Group B represents the reverse.

PAGE 103

103 Figure 4-3. Frequency of farmers knowing about management alternatives for decreasing monk parakeet damage before the interviews.

PAGE 104

104 Figure 4-4. Frequency of farmers reporting management alternatives for decreasing monk parakeet damage as the most effective one among all known alternatives.

PAGE 105

105 Table 4 1. Summary of regression results for socio psychological and sociodemographic variables and preferences for management strategies. Values for each variable correspond to the regression coefficient and their respective standard error (in parenthesis). The table includes only variables with statistically significant relationships with preferences, either i n the bivariate Chi square tests or the regression models. Significance was set as p= 0.05 for both tests. Management strategies are ordered based on farmers preferences. Management strategies: RC= reproductive control, LC= lethal control, IPM= integrated pest management CP= crop protection, AP= agricultural practices, HM= habitat management, CR= capture and relocation. Variable name RC LC IPM CP AP HM CR Socio psychological factors Perception of problems with monk parakeet Perception of damage in the last 3 yrs Perception of damage trend Tolerance to damage Importance of monk parakeet damage 0.38 (0.16) Previous knowledge of each strategy 0.94 (0.38) 1.13 (0.55)** 1.33 (0.53) 1.35 (0.52) Beliefs about effectiveness of each strategy 0.44 (0.16) 0.71 (0.20) 0. 50 (0.23)** 0.01 (0.20) 0.18 (0.26) 0.58 (0.25)** Attitudes toward monk parakeets 1.26 (0.25) 1.53 (0.29) 1.07 (0.24) 0.96 (0.24) 0.80 (0.28) Subjective norms about monk parakeet control 0.57 (0.23) 0.01 (0.01)* 0.45 (0.23)** Perceived contr ol of bird pest management 0.33 (0.23)* Socio demographic factors Age 0.34 (0.17)** 0.34 (0.17) 0.41 (0.16)** Education 0.56 (0.20) 0.44 (0.20)** 0.41 (0.20)** Farm area 0.31 (0.23)* 0.29 (0.23)* Percentage area devot ed to crops 0.18 (0.22)* Social participation 0.57 (0.24)* Statisti cally significant relationship i n the bivariate Chi square test but not in the regression model. ** Statistically significant relationship i n the regression model but not in the Chi square test.

PAGE 106

106 Table 4 2. Top performing variables related to preferences of farmers for management strategies to decrease monk parakeet damage to crops based on the AIC value and percent concordance, which represents the association of predic ted probabilities and observed responses Models with and are presented here. Management strategies are ordered based on farmerspreference. AIC values are not comparable among strategies, because they correspond to different data sets (one for each management strat egy). Method Variable name AIC Percent concordant Reproductive control (RC) Attitudes toward monk parakeets 203.93 54.40 Lethal control (LC) Attitudes toward monk parakeets 192.81 59.20 Integrated Pest Management (IPM) Previous knowledge of each strate gy Beliefs about effectiveness of each strategy 229.41 230.67 20.00 20.30 Crop protection (CP) Attitudes toward monk parakeets 222.77 49.70 Agricultural p ractices (AP) Attitudes toward monk parakeets 226.58 48.30 Habitat management (HM) Perception of da mage trend Subjective norms about monk parakeet control Tolerance to damage 218.37 218.97 220.17 45.70 42.90 37.90 Capture and relocation (CR) Attitudes toward monk parakeets Previous knowledge of each strategy 140.55 142.36 48.70 32.10

PAGE 107

107 Table 4 3. Res ults from bivariate Chi square test evaluating the correlation between independent variables. Values correspond to the Chisquare test statistic and the corresponding pvalue in parenthesis. In cases where only one value is reported, it corresponds to the p value of the Fisher test. Attitudes toward monk parakeets Subjective norms about monk parakeet control Perceived control regarding bird pest management Socio psychological factors Perception of problems with monk parakeet Perception of da mage on the last 3 yrs 0.42 2.45 (0.87) 6.16 (0.41) Perception of damage trend 0.28 5.63 (0.69) 9.68 (0.29) Tolerance to damage 0.41 1.32 (0.86) 5.62 (0.23) Importance of monk parakeet damage (loss) 7.68 (0.26) 2.00 (0.92) 6.24 (0.40) Previous knowledge of each strategy 0.62 0.21 0.54 Beliefs about effectiveness of each strategy 0.04 0.03 0.57 Socio demographic factors Age 11.91 (0.06) 6.38 (0.38) 7.06 (0.31) Education < 0.001 0.32 2.38 (0.88) Farm area 1.88 (0.76) 4.7 4 (0.31) 6.49 (0.16) Percentage area devoted to crops 5.59 (0.23) 5.75 (0.22) 3.09 (0.54) Social participation 2.35 (0.31) 0.51 (0.77) 0.38 (0.83) Indicates statistically significant relationships at p= 0.05.

PAGE 108

108 CHAPTER 5 CONCLUSIONS Influence of Lo ca l and Landscape V ariables o n Monk Parakeet Abundance or Damage in Crop Fields and N esting S ites Agricultural landscapes are a mixture of cultivated and uncultivated patches (fields), varying in composition (i.e., amount of land cover type s in the landscape) and configuration (i.e. spatial arrangement of patches within the landscape) at multiple spatial and temporal scales (Forman and Godron 1986; Burel and Baudry 1995; Holt et al. 1995; Landis and Marino 1999). Mobile species using these landscapes, includ ing b ird species causing damage to crops often use cultivated and non cultivated patches in their life cycles. The abundance and distribution of these cultivated and noncultivated patches in the agricultural landscape may influence both the abundance and damage of a bird species o n a particular patch or crop field ( Otis and Kilburn 1987; Tourenq et al. 2001; Amano et al. 2004, 2008; Hagy et al. 2008) Similarly, the abundance and distribution of these cultivated and noncultivated patches may influence the abundance of birds in nesting sites within the agricultural landscape ( Bruun and Smith 2003; Surmacki 2005). M onk parakeet abundance and damage varied in this study with the type of the crop in the field, being greater in sunflower than in corn fields ( Chapter 2). Although monk parakeet abundance and damage were cor related particularly in cor n fields, I was able to explain damage to crops by monk parakeets better than monk parakeet abundance. Probably, because damage is cumulative and abundance data rep resent an instantaneous picture, relation ships of damage to within field, field and landsca pe variables emerged more clearly than with abundance data

PAGE 109

109 L andscape characteristics around crop fields were consistently more important than local characteristic s of the crop field for explaining monk parakeet damage in a field (Chapter 2) This result may indicate the importance of landscape processes, such as landscape complementation and supplementation, for a central place forager such as the monk parakeet M o nk parakeets use nests all year around, both for breeding and roosting, and they travel limited distances each day from the nest to foraging sites (Spreyer and Bucher 1998) T he increase in damage in sunflower fields proximate to sites with manmade struct ures and trees, and in corn or sunflower fields surrounded by abundant trees, may reflect lower energetic costs for monk parakeets traveling short distances from the nest or loafing areas to foraging sites The proximity of foraging and nesting sites may r esult in larger numbers of parakeets aggregating at those sites or the lower energetic cost may favor an increase in population size of parakeets and, consequently, damage on particular plots within those landscapes. Additionally, the increase in damage in crop fields with greater availability of alternative foraging sites for monk parakeets on the landscape around the crop fields may be related to the generalist forag ing behavior of monk parakeets (Bucher et al. 1991; Hyman and Pruett Jones 1995), which ma y allow potentially higher populations in the area and, consequently, higher damage in particular fields within that area L oc al characteristics of the field, although less important than landscape characteristics, also favored monk parakeet damage to cr op fields particularly on sunflower fields. Fields with small area, low plant density and high percentage of patches with trees around the field usually were more prone to damage by monk parakeets than other sunflower fields. These results support the nee d of considering

PAGE 110

110 local and landscapelevel variables for predicting and managing bird damage to crop fields (Clergeau 1995 ; Tourenq et al. 2001; Amano et al. 2008; Hagy et al. 2008). In contrast to damage, t he density of monk parakeet nests in inhabited farms with e ucalyptus trees was not clearly explained by any variable or combination of variables modeled in this study, either at local or landscape level (Chapter 3) T he proportion of eucalyptus canopy area in the nesting patch, a local variable, was more important than any other variable at local and landscape level, but this variable had low explanatory capability. Several factors could explain t he poor performance of local and landscape variables in explaining density of monk parakeet nests S ome import ant variables may not have been measured or the way the variables were measured may not have been optimal for capturing the influence of those variables. For example, buffer extents used in this study may have been not large enough to capture habitat features related to nest settlement, such as availability of foraging sites. Finally, f actors related to the behavior of monk parakeets could have influenced nest density more than habitat variables at local and landscape levels including colonial habits nata l philopatry, or generalist foraging behavior of monk parakeets T he importance of landscape variables for explaining monk parakeet damage in foraging sites (crop fields) compared to monk parakeet abundance in nesting sites (sites with eucalyptus trees) m ay be explained by the range of movements associated with each process. The scale of response to the environment by mobile species depends on the ecological process under consideration and the movement ranges of these species (Addicott et al. 1987 ; Wiens 1 989 ; Ho lland et al. 2004). Daily movements of monk parakeets from the nest to feeding area are generally between 3 and 5 km while

PAGE 111

111 breeding, but can reach 24 km outside the reproductive season (Spreyer and Bucher 1998). Dispersal distances from natal nest t o first breeding site usually ranges between 0. 3 to 2 km and when changing nests from one year to another, mean distance of movement is 0.5 k m ( maximum= 2 5 km, Martin and Bucher 1993; Spreyer and Bucher 1998). Therefore, daily foraging activities of monk parakeets comprise a bigger area than annual reproductive activities Consequently, landscape scale variables may influence foraging activities more than reproductive activities. Human D imensions of Conflicts between M onk Parakeets and Crop P roduction Po pulation control strategies, such as nest destruction and killing of birds, were perceived by farmers as the most effective strategies for decreasing monk parakeet damage to crops and also were the most preferred (Chapter 4) The use of these strategies is historical in Argentina (Bucher and Bedano 1976; Bucher 1984; Arambur 1991; Bucher 1992 a and b ; Bruggers et al. 1998), including t he region of this study (Zaccagnini and Bucher 1983 ; Gimnez and Salomn 2000). The stated preferences were consistent with field observations from my assessment of factors influencing the distribution of parakeet nests. I n farms where farmers used control methods against monk parakeets, the mos t common methods were centered on population control, applying lethal control by shooting, either as the only control measure or combined with removing or burning nests. Nonlethal population control measures were centered in reproductive control, including removing nests burning nests, or capturing nestlings alive for pet trade. Pr efer ences of farmers for management strategies were related more strongly to attitudes toward monk parakeets than to any other sociopsychological or sociodemographic factor. Similar to what has been found in previous studies looking at the

PAGE 112

112 influence of attit udes on preferences for management actions involving wildlife species (e.g., Bjerke et al. 1998 ; Stout et al. 1997 ; Don Carlos et al. 2009; Loyd and Miller 2010), most farmers with negative attitudes toward monk parakeet preferred invasive population contr ol methods, such as lethal and reproductive control, and farmers with positive attitudes toward monk parakeets preferred nonlethal strategies, such as crop protection and agricultural practices. Given crop damage from monk parakeets was considered tolerable, the predominantly negative attitudes toward monk parakeets could be related to past problems more than the perception of actual damage (Zinn and Andelt 1999). This proposition may be supported by the relationship of attitudes toward monk parakeets with age, with older farmers having predominantly negative attitudes compared to younger farmers. Preferences of farmers for management strategies also were strongly associated with perceived efficacy of management strategies and, to a lesser extent, to previ ous knowledge about those strategies. Other socio psychological factors, such as subjective norms and perceived control, were related to preferences of farmers for management strategies but to less er degree than attitudes. Finally, p references for management strategies, including lethal or reproductive control, generally were not related to perceptions of magnitude of damage by monk parakeets or socio demographic variables, such as age and education, al though there were some disparities depending on the management strategy under consideration. The prediction of behavior or intention to act is a complex process based on multiple factors (Fishbein and Ajzen 1975; Hines et al. 1986; Ajzen 1991; Norton and Mumford 1993). Socio psychological variables, such as at titudes, perception of efficacy

PAGE 113

113 economic orientation, etc., and cognitive variables, such as knowledge of environmental problems and how to take action, are known to influence the intention to act (or verbal commitment) in responsible environmental behavi ors more than socio demographic factors such as age, income, education and gender (Hines et al. 1986). Results from this study generally were consistent with previous studies looking at human dimensions of human wildlife conflicts (e.g., Bjerke et al. 199 8 ; Stout et al. 1997; Don Carlos et al. 2009; Loyd and Miller 2010) and form a basis to understand farmers preferences for current population control measures. Management I mplications Monk parakeet damage to corn and sunflower fields has been documented as an important concern for agricultural producers in some areas of Argentina (Bucher 1984, 1992a, b; Bruggers y Zaccagnini 1994; Bruggers et al. 1998). In my study, far mers perceived a problem exists and indicated a preference for population control methods, either lethal (e.g., tox ic baits nest poisoning with pesticides ) or reproductive ( e.g., nest removal, burning, etc.), to reduce monk parakeet damage to crops However, damage in corn and sunflower fields evaluated in this study w as relatively low (< 5 % of damaged plants, corresponding to less than 3% of grain loss, Canavelli et al. 2008), and damages were perceived as low to moderate but tolerable by most farmers Additionally, no control measures against monk parakeet damage were observed on any of th e surveyed crop fields in this study, and people did not appl y any control measures against monk parakeets a t most farms with eucalyptus trees surveyed in this study Therefore, there is little evidence from this study for the need of control ling monk parakeets at least in the region where th e study was conducted.

PAGE 114

114 Based on this study, when management measures are needed, managers sh ould consider local factors, such c rop type, field area and plant density, when planning management measures to prevent monk parakeet damage to crop fields, particularly to sunflower fields. Also managers trying to reduce t he number of monk parakeet nests in inhabited farm s with eucalyptus trees would have to consider limiting the available nest sites at local scales This coul d be done by removing eucalyptus trees or by modifying the structure of the eucalyptus trees. Potentially, the removed eucalyptus trees could be replaced by other trees not adequate for nesting. However, none of these management strategies has been evaluat ed. The only management technique at the landscape scale that emerged from this study that might decrease damage to crops fields by monk parakeets and abundance of nests in farms would be to decrease the amount of tree patches around crop fields or farms For example, farmers could try to locate crop fi elds at least 1 km from places with manmade structures and trees (cascos or farms) and other patches with trees in order to prevent monk parakeet damage. However, decreasing the amount of tree patches around the crop fields or the farms will be difficult without cutting down the trees. E liminating trees, either for decreasing monk parakeet damage to crops or nest abundance in inhabited farms, would result in a more simplified landscape, which could have imp ortant consequences for biodiversity and ecosystem services provided by biodiversity, including the regulation of other crop pests ( Bianchi et al. 2006; Power 2010; Batry et al. 2011). Detailed information about the cost of control not only economic al but also environmental, compared to economic loss from damage would be particularly useful to put the problem of damage to crops by monk parakeets in

PAGE 115

115 perspective for farmers and to select management measures that are socially optimal ( i.e., with more benefit s for the whole society than for individual landow ners, Tisdell 1982) Se veral management strategies are currently available at field level that may reduce monk parakeet damage to crop fields or protect those fields from damage, including altering agricul tural practices and crop protection measures such as increasing crop density and sowing deterrent crops, or using bior epellents (Canavelli and Arambur, in press ). Unfortunately, no evaluations have been made o f the efficacy of these management strategies to decrease monk parakeet damage to crops or the cost benefit ratio of applying these techniques Additionally, this study indicated that these management strategies currently are not among the most preferred by farmers Considering reproductive control was the most preferred management strategy by farmers in this study, a population control that could be promissory for the future is chemical contraception. C urrent research indicates the effectiveness of Diazacon as a chemical inhibitor of reproduction for monk parakeets (Avery et al. 2008). However, at the moment there is not a registered product for reducing fertility of monk parakeets in any country worldwide. In addition, multiple aspects related to the use of chemical contraceptives, such as biologic al feasibility, economic practicality and health and safety issues, including impact on nontarget species (Avery et al. 2008 ; Fagerstone et al. 2010), would have to be addressed bef ore a product is available for reproductive control of monk parakeets in A rgentina. In order to shape farmers opi nions and management decisions about strategies other than lethal or reproductive control i n the way they are currently applied, it is

PAGE 116

116 n ecessary to evaluate the efficacy of alternative strategies, both at local and l andscape levels, in those regions where conflicts with monk parakeets are important. G iven the current uncertainties i n the outcome of management actions to decrease monk parakeet damage to crops, an adaptive management approach, in which multiple land owners are involved and experiments with different management options are conducted in the region ( Holling 1978; Walters 1997; Shea et al. 2002; Parkes et al. 2006) is appropriate (Canavelli and Zaccagnini 2007; Canavelli and Arambur, in press ). E xtension ac tions within a management program also would have to integrate people with different points of view, in order to avoid public controversies about management. Finally, research and e xtension actions should be pr oactive, so that intolerable losses are antici pated and avoided instead of trying to eliminate a situation once it has occurred.

PAGE 117

117 APPENDIX A COMPLEMENTARY RESULTS FROM CHAPTER S 2 AND 3

PAGE 118

118 Table A 1. Correlations between landscape metrics used to quantify landscape composition and configuration aroun d focal fields Values are given for the 3000m buffer for sunflower fields. Different buffer widths for different crops (corn or sunflower) have different values for the correlation coefficients, but the relationships between variables are qualitatively s imilar. CRPLAND* CRSHAPE CRMNN CRCLUMP TRPLAND TRSHAPE TRMNN TRCLUMP PSTPLAND CRPLAND 1.00 0.63 0.71 0.38 0.61 0.52 0.52 0.22 0.26 CRSHAPE 1.00 0.63 0 65 0.47 0.27 0.45 0.32 0.41 CRMNN 1.00 0.79 0.41 0.32 0.26 0.19 0.41 CRCLUMP 1.00 0 .39 0.32 0.26 0.30 0.45 TRPLAND 1.00 0.89 0.70 0.64 0.34 TRSHAPE 1.00 0.65 0.54 0.39 TRMNN 1.00 0.25 0.14 TRCLUMP 1.00 0.03 PSTPLAND 1.00 CRPLAND = Percentage of landscape comprised by crops susceptible to damage (corn and sunflower), CRSHAPE = Mean shape of crop patches susceptible to damage in the landscape, CRMNN = Mean nearest neighbor distance among crop patches susceptible to damage, considering all patches on the landscape, CRCLUMP = Clumpiness index of c rop patches susceptible to damage, TRPLAND = Percentage of landscape comprised by tree patches, TRSHAPE = Mean shape of tree patches in the landscape, TRMNN = Mean nearest neighbor among tree patches, considering all tree patches on the landscape, TRCLUMP = Clumpiness index of tree patches, PSTPLAND = Percentage of landscape comprised by pastures and other agricultural uses (including weedy and fallow fields)

PAGE 119

119 Table A 2 Suite of models used to describe the relative abundance and damage of monk parakeet to crop fields at within field, field and landscape levels in Paran (Entre Ros, Argentina), 20072008 Models were run for corn and sunflower separately. Within field Field Landscape 2 Within field PLTDEN PHENST WDCOV AREA SHAPE TREES CRPLAN CRCLUMP TR PLAND PSTPLAND 1 x 2 x 3 x 4 x x 5 x x 6 x x 7 x x x Field 1 x 2 x 3 x 4 x x 5 x x 6 x x 7 x x x Landscape 2 1 2 x 3 x 4 x 5 x 6 x x 7 x x 8 x x 9 x x 10 x x x 11 x x x

PAGE 120

120 Table A 2. Continued Within field Field Landscape 2 Multi level 3 PLTDEN PHENST WD COV AREA SHAPE TREES CRPLAN CRCLUMP TRPLAND PSTPLAND Sunflower 1 x x 1x 2 x x 3x 3 x x 5x Corn 1 x x 1x 2 x x 3x 3 x x 5x 1 PLTDEN = Plant density, PHENST = Phenological stage, WDCOV = Weed coverage, AREA = Field area, SHAPE = field shape index, TREES = Abundance of t rees on border, CRPLAND = Percentage of crops susceptible to damage (corn and sunflower), CRCLUMP = Clumpiness index of crop patches susceptible to damage, TRPLAND = Percentage of tree patches, PSTPLAND = Percentage of pastures and other agricultural uses (including weedy and fallow fields) 2 Models at this level were estimated at three buffer extents (1, 3 and 5 km) around each crop field. Models at 1 km buffer extent included an additional variable (DISTCO= d istance to the nearest site with manmade structures and t rees ). 3 Multi level models were estimated only for damage. N umbers preceding the letter code x for landscape variables indicate buffer sizes (in kilometers)

PAGE 121

121 Table A 3 Correlations between landscape metrics used to quantify landscape composition and configuration around nesting patches. Values are given for the 3000 m buffer. Different buffers ha d different values f or the correlation coefficients, but the r elationships between vari ables we re qualitatively similar. CRPLAND PSTPLAND TRPLAND TRPD TRED TRMNN TRCLUMP TRSHAPE CRPLAND 1.00 0.13 0.35 0.54 0.10 0.28 0.37 0.39 PSTPLAND 1.00 0. 59 0.42 0.46 0.51 0.53 0.53 TRPLAND 1.00 0.25 0.87 0.93 0.6 8 0.86 TRPD 1.00 0.07 0.18 0.25 0.32 TRED 1.00 0.89 0.44 0.82 TRMNN 1.00 0.59 0.87 TRCLUMP 1.00 0.54 TRSHAPE 1.00 CRPLAND= Percentage of landscape with preferred food crops PSTPLAND= Percentage of landscape with pastures and other agricultural uses (including weedy and fallow fields), TRPLAND= Percentage of landscape with trees (either native or introduced) in the landscape TRPD= Density of patches with trees on the landscape (#/km2) ,TRED= Edge density of tree patches, TRMNN = Mean nearest neighbor among all tree patches in the landscape, TRCLUMP = Clumpiness index of tree patches, TRSHAPE = Shape index for tree patches

PAGE 122

122 APPENDIX B QUESTIONNAIRE STRUCTURE AND CONTENT A.1. Farmer and farm identification The first page of the questionnaire served to obtain personal information about the farmer (name, address, phone number, etc.) and the farm to which the farmer was referring the answers, including its geographical reference. A.2. Farmers perceptions and beliefs abou t monk parakeet damage to crops a. Perception of damage on the last three years, the last crop season (August 2006May 2007) and its trend on the last 3 years. b. Tolerance to damage on specific crops: corn, sunflower, soybean, wheat, sorghum, alfalfa, foxtail and cattail millet. c. Importance of monk parakeet damage compared to other crop loss causes, including insects, weeds, diseases and weather. d. Factors farmers consider could favor monk parakeet damage to crops, including proximity to woodland, utility tower, agricultural practices, etc. A.3. Farmers perceptions and attitudes toward monk parakeets a. Perception of monk parakeet abundance and its trend on the last 3 years. b. Attitudes toward monk parakeets (Likert scale). c. Opinion of farmers about different types of woodlands as refuge for monk parakeets. A.4. Farmers knowledge and preferences for management strategies to decrease monk parakeet damage to crops a. Knowledge of at least one management strategy. b. Knowledge about particular management strategies (7 options).

PAGE 123

123 c. Beliefs about effectiveness of particular management strategies. d. Preferences for management strategies (paired comparisons). e. Willingness to try new management alternatives. f. External factors influencing the decision to use a particular management strategy including economic cost, toxicity and available information. A.5. Personal and external influences on the application of management strategies a. Personal confidence about particular techniques b. External factors limiting the application of these techniques c. Farmers opinion on what they suppose other people expect for them to do in relation to monk parakeet control. d. Influence of other peoples opinion on farmers management decisions. A.6. Socio demographic information Socio demographic variables, including age, educational level, operated area, area of farm devoted to crops and other productive activities, percentage of income from crops and other productive activities, affiliation of a farmer to a farming or a conservation organization, participation i n fa rmers meetings, and information sources about pest management. F or a copy of the questionnaire in Spanish, see Appendix C.

PAGE 124

124 APPENDIX C QUESTIONNAIRE IN SPANISH EVALUACIN DE PROBLEMAS OCASIONADOS POR COTORRAS EN CULTIVOS DEPARTAMENTO PARAN Cuestionario para entrevistas personales a productores Fecha: ____ / ____ / ____ (da/mes/ao) 1. IDENTIFICACIN DEL PRODUCTOR 1.1. Apellido y nombre del productor __________________ ___________________________________________________________ 1.2. Domicilio Calle/Ruta, N/km (solo si difiere de la direccin de la explotacin) _____________________________________________________________________________ Cdigo Postal: ____________________ Localidad o paraje ms cercano: __________________________________________________ Tel fono: _______________________ Celular: ______________________ ___ Correo electrnico: _____________________________________________________________ 2. IDENTIFICACIN DE LA EXPLOTACIN AGROPECUARIA 2.1. Nombre:__________________________________________________________________ 2.2. Ubicacin: Departamento: Paran ____ Otro________________________ Calle/Ruta, N/km (Nomenclatura catastral) Cdigo Postal: ____________________ Localidad o paraje ms cercano: __________________________________________________ Ubicacin respecto al mismo (km, orientacin):_______________________________________ _____________________________________________________________________________ Coordenadas geogrficas: Latitud: __________________ S Longitud: _________________ W

PAGE 125

125 3 INFORMACION SOBRE DAO POR COTORRAS Q1 Ha experimentado dao por cotorras en cultivos de su produccin en los ltimos 3 aos (PERIODO DE REFERENCIA: MAYO 200 4 MAYO 2007) s i o no ? NO.. 1 IR A Q8 SI........ 2 NO CONTESTA. 9 Q2 El dao de cotorras ha sido insignificante, moderado, intenso o total en estos 3 aos? INSIGNIFICANTE 1 MODERADO 2 INTENSO 3 TOTAL 4 INDECISO 7 NO CONTE STA 9 Q3 Si tuviera que asignar un porcentaje al dao total por cotorras en esta ltima campaa (20067 PERODO DE REFERENCIA: 1 AGOSTO 2006 31 MAYO 2007) cunto estimara que fue el mismo? < 5% 1 5 10% 2 10 25% 3 25 50% 4 50 75% 5 75 100% 6 I NDECISO 7 NO CONTESTA 9 Q4 Este porcentaje sera menor, igual o mayor al dao promedio de los ltimos 3 aos? MENOR 1 IGUAL 2 MAYOR 3 INDECISO 7 NO CONTESTA 9 Q5 En base a su experiencia, considerara el dao por cotorras tolerable o intolerable en los siguientes cultivos? NO HACE EL CULTIVO HACE CULTIVO INDECISO ID CULTIVO NO DAO DAO TOLERABLE DAO INTOLERABLE 1 MAZ 0 1 2 3 7 2 GIRASOL 0 1 2 3 7 3 SOJA 0 1 2 3 7 4 TRIGO 0 1 2 3 7 5 SORGO 0 1 2 3 7 6 ALFALFA 0 1 2 3 7 7 MOHA 0 1 2 3 7 8 MIJO 0 1 2 3 7 9 OTRO (ESPECIFICAR) __________________ 0 1 2 3 7 NO CONTESTA 9

PAGE 126

126 Q6 Las prdidas ocasionadas por cotorras sera n menor, igual o mayor que la producidas por insectos, malezas, enfermedades, clima, cosechadora, u otras causa s ? ID MENOR IGUAL MAYOR INDECISO NO CONTESTA 1 INSECTOS 1 2 3 7 9 2 MALEZAS 1 2 3 7 9 3 ENFERMEDADES 1 2 3 7 9 4 CLIMA 1 2 3 7 9 5 COSECHADORA 1 2 3 7 9 6 OTRAS(MENCIONAR) _______________ 1 2 3 7 9 Q7 A continuacin, voy a mencionarle factores que podran favorecer el dao por cotorras a sus cultivos. En su opinin, cules seran factores importantes en su explotacin ? No importante Algo importante Muy importante INDECISO NO CONTESTA 1 ABUNDANCIA DE COTORRAS 0 1 2 7 9 2 PROXIMIDAD A MO NTES 0 1 2 7 9 3 PROXIMIDAD A TORRES DE ALTA TENSIN 0 1 2 7 9 4 PRCTICAS AGRCOLAS (SIEMBRA, COSECHA, ETC.) 0 1 2 7 9 5 ALIMENTO DISPONIBLE DURANTE EL INVIERNO (FEEDLOTS, RASTROJO, ETC) 0 1 2 7 9 6 OTROS (ESPECIFICAR) ________________________________ ___________ 0 1 2 7 9 4 INFORMACIN SOBRE COTORRAS Q8 Piensa Ud. que en su explotacin hay pocas cotorras, algunas, o muchas? POCAS 1 ALGUNAS 2 MUCHAS 3 INDECISO 7 NO CONTESTA 9 Q9 Si compara este nmero con el nmero promedio de cotorras en lo s ltimos 3 aos, pensara que es menor, igual o mayor? MENOR 1 IGUAL 2 MAYOR 3 INDECISO 7 NO CONTESTA 9 Q1 0 A continuacin nos gustara conocer su opinin acerca de las cotorras ( en este momento y en general, no necesariamente asociado al moment o en que hacen o no hacen dao) Por favor, le agradecera nos indique cunto est en acuerdo o desacuerdo con las frases siguientes:

PAGE 127

1 27 En desacuerdo De acuerdo Indeciso No contesta 1 Las cotorras nos alegran con su presencia. 1 2 7 9 2 Afortunadame nte, podemos convivir con las cotorras. 1 2 7 9 3 Las cotorras son criaturas que deben ser protegidas. 1 2 7 9 4 Me molestan las cotorras porque disminuyen mi produccin. 1 2 7 9 5 Las cotorras me imposibilitan la siembra de girasol o maz. 1 2 7 9 6 D etesto el parloteo de las cotorras. 1 2 7 9 7 Las cotorras no sirven para nada. 1 2 7 9 8 Me gustan las cotorras. 1 2 7 9 9 Me molestan las personas que protegen a las cotorras. 1 2 7 9 10 Me gusta observar a las cotorras en el campo. 1 2 7 9 11 Me du ele ver como se persigue a las cotorras. 1 2 7 9 12 Valoro mucho a las cotorras. 1 2 7 9 13 Las cotorras solo me hacen perder dinero. 1 2 7 9 14 Odio a las cotorras. 1 2 7 9 Q1 1 A continuacin, voy a mencionarle distintos tipos de monte que podran ac tuar como refugio para las cotorras P or favor p odra indicarme en cada caso si Ud. considera que los mismos actan como un refugio para las cotorras o no ? REFUGIO? ID TIPO DE MONTE NO SI INDECISO 1 Nativo 1 2 7 2 De eucaliptus 1 2 7 3 Introducid o distinto de eucaliptus 1 2 7 4 Mixto con eucaliptus 1 2 7 5 Mixto sin eucaliptus 1 2 7 NO CONTESTA 9 5. INFORMACIN SOBRE MANEJO DEL DAO POR COTORRAS Q12 Conoce algn mtodo para disminuir el dao por cotorras si o no ? NO.. 1 IR A Q 19 SI........ 2 NO CONTESTA. 9 Q13 A continuacin, voy a leer le un listado de 7 mtodos que son comnmente utilizados en el manejo de aves perjudiciales. Para cada uno, podra por favor indicarme si lo conoce o no, y por qu medio lo conoce (experiencia propia, experiencia de los vecinos, u otros medios)?

PAGE 128

128 CONOCE MEDIO METODO NO SI Experiencia propia Vecinos Otros (especificar) 1 Control letal (como trampas, veneno en los nidos, cebos txicos, disparos de escopeta) 1 2 3 4 5 ____________ 2 Proteccin del cultivo (como repelentes qumicos, auditivos caones de explosin, disparos de escopeta, visuales) 1 2 3 4 5 ____________ 3 Prcticas agrcolas (siembra profunda, cosecha anticipada) 1 2 3 4 5 ____________ 4 Manejo del ambient e (poda o eliminacin de rboles, siembra de cultivos trampa) 1 2 3 4 5 ____________ 5 Control de reproduccin (como volteo y /o quema de nidos, aceite en los huevos, sustancias qumicas que evitan la reproduccin) 1 2 3 4 5 ____________ 6 Trampeo (captu ra viva) y/o reubicacin 1 2 3 4 5 ____________ 7 Manejo Integrado de Aves Plaga (MIP) 1 2 3 4 5 ____________ Q14 Para cada uno de los mtodos conocidos por Ud., ya sea por experiencia propia o por otros medios, podra indicarme por favor si considera q ue no es efectivo o es efectivo ( algo o muy efectivo ) ?. A continuacin, voy a leerle la lista nuevamente. METODO NO EFECTIVO ALGO EFECTIVO MUY EFECTIVO SIN OPININ 1 CONTROL LETAL 1 2 3 0 2 PROTECCIN DEL CULTIVO 1 2 3 0 3 PRCTICAS AGRCOLAS 1 2 3 0 4 MANEJO DEL AMBIENTE 1 2 3 0 5 CONTROL DE REPRODUCCION 1 2 3 0 6 TRAMPEO Y/O REUBICACION 1 2 3 0 7 MANEJO INTEGRADO DE PLAGAS 1 2 3 0 Q15 Ahora, en base a la efectividad que acabamos de mencionar, cul piensa que es el ms efectivo?. Si lo dese a, con gusto puedo leerle la lista nuevamente. 1. MAS EFECTIVO 2. SEGUNDO MAS EFECTIVO 3. TERCERO MAS EFECTIVO Q16 A continuacin, vamos a presentarle pares de alternativas para disminuir el dao por cotorras. Para cada par, le agradeceramos no s indique cul de las 2 alternativas prefera aplicar, asumiendo fuera necesario disminuir el dao por cotorras a sus cultivos (por ejemplo en la prxima campaa).

PAGE 129

129 ID# Id1 Alternativa 1 Id2 Alternativa 2 1 1 Control letal 2 Proteccin del cultivo 2 3 Prcticas agrcolas 1 Control letal 3 1 Control letal 4 Manejo del ambiente 4 5 Control de reproduccin 1 Control letal 5 1 Control letal 6 Captura viva 6 7 Manejo Integrado de Plagas 1 Control letal 7 2 Proteccin del cultivo 3 Prcticas agrcolas 8 4 Manejo del ambiente 2 Proteccin del cultivo 9 2 Proteccin del cultivo 5 Control de reproduccin 10 6 Captura viva 2 Proteccin del cultivo 11 2 Proteccin del cultivo 7 Manejo Integrado de Plagas 12 4 Manejo del ambiente 3 Prcticas agrcolas 13 3 Prcticas agrcolas 5 Control de reproduccin 14 6 Captura viva 3 Prcticas agrcolas 15 3 Prcticas agrcolas 7 Manejo Integrado de Plagas 16 4 Manejo del ambiente 5 Control de reproduccin 17 6 Captura viva 4 Manejo del ambiente 18 4 Manejo del a mbiente 7 Manejo Integrado de Plagas 19 5 Control de reproduccin 6 Captura viva 20 7 Manejo Integrado de Plagas 5 Control de reproduccin 21 6 Captura viva 7 Manejo Integrado de Plagas 22 5 Control de reproduccin 2 Proteccin del cultivo 23 3 Prcti cas agrcolas 4 Manejo del ambiente Q17 Nuevamente, asumiendo que fuera necesario para Ud. disminuir el dao por cotorras a sus cultivos (en la prxima campaa, por ejemplo) estara o no dispuesto a probar tcnicas de manejo diferentes a las que conoce? NO.. 1 IR A Q21 SI........ 2 NO CONTESTA. 9 Q18 De ser as, cules alternativas estara dispuesto o no estara dispuesto a probar ? METODO NO DISPUESTO DISPUESTO INDECISO NO CONTESTA 1 CONTROL LETAL 1 2 7 9 2 PROTECCIN DEL CUL TIVO 1 2 7 9 3 PRCTICAS AGRCOLAS 1 2 7 9 4 MANEJO DEL AMBIENTE 1 2 7 9 5 CONTROL DE REPRODUCCION 1 2 7 9 6 TRAMPEO Y/O REUBICACION 1 2 7 9 7 MANEJO INTEGRADO DE PLAGAS 1 2 7 9 IR A Q23 Q19 En este caso, le gustara conocer o no alternativas de manejo para disminuir el dao por cotorras a sus cultivos? NO.. 1 IR A Q21 SI........ 2 NO CONTESTA. 9

PAGE 130

130 Q20 A continuacin, le voy a mencionar alternativas de manejo que son comnmente utilizadas para disminuir los daos por aves perjud iciales. Las voy a mencionar en pares (es decir, de a dos) y, para cada par, le agradecera me indique cul de los 2 alternativas deseara conocer primero. Si el mtodo no es totalmente claro, con gusto puedo brindarle mayores detalles. ID# Id1 Alternati va 1 Id2 Alternativa 2 1 1 Control letal 2 Proteccin del cultivo 2 3 Prcticas agrcolas 1 Control letal 3 1 Control letal 4 Manejo del ambiente 4 5 Control de reproduccin 1 Control letal 5 1 Control letal 6 Captura viva 6 7 Manejo Integrado de Pla gas 1 Control letal 7 2 Proteccin del cultivo 3 Prcticas agrcolas 8 4 Manejo del ambiente 2 Proteccin del cultivo 9 2 Proteccin del cultivo 5 Control de reproduccin 10 6 Captura viva 2 Proteccin del cultivo 11 2 Proteccin del cultivo 7 Manejo Integrado de Plagas 12 4 Manejo del ambiente 3 Prcticas agrcolas 13 3 Prcticas agrcolas 5 Control de reproduccin 14 6 Captura viva 3 Prcticas agrcolas 15 3 Prcticas agrcolas 7 Manejo Integrado de Plagas 16 4 Manejo del ambiente 5 Control de r eproduccin 17 6 Captura viva 4 Manejo del ambiente 18 4 Manejo del ambiente 7 Manejo Integrado de Plagas 19 5 Control de reproduccin 6 Captura viva 20 7 Manejo Integrado de Plagas 5 Control de reproduccin 21 6 Captura viva 7 Manejo Integrado de Pla gas 22 5 Control de reproduccin 2 Proteccin del cultivo 23 3 Prcticas agrcolas 4 Manejo del ambiente IR A Q23 Q21 Por favor, sera tan amable de indicarme al menos una razn de su negativa? NO CONTESTA 99 Q22 Asumiendo que en la prxima campaa observa daos por cotorras en alguno de sus cultivos, podra indicarme cul de los siguientes mtodos probablemente aplicara? Si el mtodo no es totalmente claro, con gusto puedo brindarle mayores detalles.

PAGE 131

131 APLICARIA? METODO NO SI I NDECISO NO CONTESTA 1 Control letal (como trampas, veneno en los nidos, cebos txicos, disparos de escopeta) 1 2 7 9 2 Proteccin del cultivo (como repelentes qumicos, auditivos caones de explosi n, disparos de escopeta, visuales ) 1 2 7 9 3 Prctica s agrcolas (siembra profunda, cosecha anticipada) 1 2 7 9 4 Manejo del ambiente (poda o eliminacin de rboles, siembra de cultivos trampa) 1 2 7 9 5 Control de reproduccin (como volteo y /o quema de nidos, aceite en los huevos, sustancias qumicas que evitan la reproduccin ) 1 2 7 9 6 Trampeo (captura viva) y/o reubicacin 1 2 7 9 7 Manejo Integrado de Aves Plaga (MIP) 1 2 7 9 CONTINUAR CON Q23 ================================================================================ Q23 A continuacin, le voy a leer un listado de 9 factores que podran influir en su decisin de aplicar una alternativa de manejo del dao por cotorras. Para cada uno, podra por favor indicarme cunto influye el factor (nada, algo o mucho) en su decisin? FACTOR NADA ALGO M UCHO INDECISO NO CONTESTA 1 Costo econmico 1 2 3 7 9 2 Costo de esfuerzo 1 2 3 7 9 3 Toxicidad 1 2 3 7 9 4 Efectividad esperada 1 2 3 7 9 5 Impacto en especies distintas a las cotorras 1 2 3 7 9 6 Requerimiento de colaboracin con vecinos 1 2 3 7 9 7 Informacin disponible 1 2 3 7 9 8 Disponibilidad en el mercado de productos 1 2 3 7 9 9 Experiencia personal o de vecinos 1 2 3 7 9 6 INFORMACIN SOBRE HABILIDADES PARA APLICAR TECNICAS DE MANEJO Q2 4 En las siguientes preguntas, imagine que se le presenta la posibilidad de aplicar una tcnica de manejo para disminuir el dao por cotorras. No interesa lo que la tcnica hace en s misma, slo que su objetivo es disminuir el dao, y es la primera vez que la utiliza. A continuacin, le mencionar u na serie de habilidades necesarias para aplicar tcnicas de manejo de conflictos con aves. Para cada una, le agradecera me indique si se siente inseguro, algo o muy seguro de poder aplicar esta nueva tcnica de manejo en funcin de la habilidad mencionada HABILIDAD INSEGUR O ALGO SEGURO MUY SEGUR O INDECISO NO CONTEST A 1 Leer, comprender y seguir instrucciones de guas de uso 1 2 3 7 9 2 Seguir estrictamente procedimientos de seguridad 1 2 3 7 9 3 Controlar maquinaria 1 2 3 7 9 4 Seleccionar la cl ase de herramientas para hacer el trabajo. 1 2 3 7 9

PAGE 132

132 Q 2 5 La habilidades antes mencionadas estn condicionadas, en algunos casos, por limitantes que impiden su aplicacin. A continuacin, voy a leer un a serie de limitantes que, en ciertos casos, podran afectar su habilidad para aplicar medidas de manejo de av es perjudiciales. Por favor, le agradecera me indique en cada caso, si considera que las siguientes limitantes no limitan o limitan (algo o mucho) dicha habilidad LIMITANTES NO LIMITAN LIMITAN ALGO LIMITAN MUCHO INDECIS O NO CONTESTA 1 Acceso restri ngido a i nformacin tcnica. 1 2 3 7 9 2 Alta complejidad de las tcnicas actuales. 1 2 3 7 9 3 Prejuicios en la comunidad sobre control letal. 1 2 3 7 9 4 Cuantificacin dificultosa de los costos y beneficios de la aplicacin de una tcnica. 1 2 3 7 9 5 A cceso restringido a maquinarias o instrumentos para el control 1 2 3 7 9 6 Condiciones ambientales adversas 1 2 3 7 9 7 I nformacin escasa sobre manejo de agroqumicos. 1 2 3 7 9 8 Alto c osto de las tcnicas actuales. 1 2 3 7 9 HABILIDAD INSEGUR O ALGO SEGURO MUY SEGUR O INDECISO NO CONTEST A 5 Considerar los costos y beneficios relativos de posi bles acciones 1 2 3 7 9 6 Tomar decisiones en base a informacin tcnica. 1 2 3 7 9 7 Estar informado sobre tcnicas de control de aves ms all de las qumicas 1 2 3 7 9 8 Saber identificar las aves involucradas en el dao. 1 2 3 7 9 9 Conocer el comportamiento de las aves involucradas en el dao. 1 2 3 7 9 10 Integrar informac in de los alrededores del campo en el anlisis del problema. 1 2 3 7 9 11 Conocer las regulaciones vigentes sobre manejo de aves silvestres. 1 2 3 7 9 12 Conocer mtodos de captura de aves (como trampas). 1 2 3 7 9 13 Conocer los plaguicidas y sus efec tos m s all de la plaga 1 2 3 7 9 14 Aplicar agroqumicos de acuerdo a las regulaciones vigentes. 1 2 3 7 9 15 Conocer propiedades de los agroqumicos 1 2 3 7 9 16 Ejecutar acciones de seguimiento (monit oreo) de las tcnicas aplicadas 1 2 3 7 9

PAGE 133

133 7 INFORMACIN SOBRE INFLUENCIAS EXTERNAS Q2 6 A continuacin, voy a mencionarle personas o grupos de personas que pueden influir en las decisiones de manejo de plagas. En cada caso, le agradecera me indique cul sera su opinin sobre lo que cada persona o grupo de per sona espera que Ud. haga para el manejo de las cotorras? No controlar nunca Controlar ocasionalmente (una vez cada 2 aos) Controlar frecuentemente (al menos 1 vez al ao) Controlar muy frecuentemente (en cada estacin) INDECISO NO CONTESTA/ NO APLICA 1 Vecinos 0 1 2 3 7 9 2 Esposa/o (de tenerla/o) 0 1 2 3 7 9 3 Extensionistas de una cooperativa 0 1 2 3 7 9 4 Agentes del gobierno nacional 0 1 2 3 7 9 5 Agentes de venta de las agroqumicas 0 1 2 3 7 9 6 Agentes del gobierno provincial 0 1 2 3 7 9 Q2 7 Cunto le preocupa (nada, algo o mucho) lo que cada persona o grupo piense sobre lo que Ud. tendra que hacer para manejar los problemas con las cotorras? No me preocupa Me preocupa algo Me preocupa mucho INDECISO NO CONTESTA/ NO APLICA 1 Vecin os 0 1 2 7 9 2 Esposa/o (de tenerla/o) 0 1 2 7 9 3 Extensionistas de una cooperativa 0 1 2 7 9 4 Agentes del gobierno nacional 0 1 2 7 9 5 Agentes de venta de las agroqumicas 0 1 2 7 9 6 Agentes del gobierno provincial 0 1 2 7 9 8 INFORMACIN S OBRE OBJETIVOS DE PRODUCCION Q 28 Como productor, asumimos tiene determinados objetivos en su produccin. En su opinin, cules de los siguientes objetivos seran importantes, desde el ms al menos importante? 1. MAS IMPORTANTE 2. SEGUNDO 3. TERCERO 9 INFORMACIN SOCIO ECONOMICA OBJETIVO CODIGO Minimiz ar riesgos 1 Minimizar costos por unidad de produccin 2 Maximizar la produccin 3 Minimizar el impacto ambiental 4

PAGE 134

134 Q 29 A continuacin, necesitaramos contar con informacin personal adiciona l. Por favor, podra indicarnos su gnero y su edad? Q3 0 Cul es el grado ms alto que complet en sus estu dios? PRIMARIA 1 SECUNDARIA 2 TERCIARIA (2 3 aos) 3 UNIVERSITARIA (4 o ms aos) 4 POSGRADO (Maestra y/o Doctorado) 5 NO CONTESTA 9 Q3 1 Por favor, podra indicarme cul es la superficie, en hectreas, de la explotacin agropecuaria de referencia de la presente entrevista (incluye tierras propias y arrendadas, actividades agrcolas y otras ) ? ______________________________ has Q3 2 Podra indicarme, a continuacin, c untas hectreas (o porcentaje aproximado del total) destin, en esta ltima campaa (20067 CORTE : 31 MAYO 2007), a la produccin de cultivos? ______________________________ has 1.GENERO FEMENINO 1 MASCULINO 2 2. EDAD <18 1 18 29 2 30 39 3 40 49 4 50 59 5 59 65 6 >65 7 NO CONTESTA 9 EN HAS < 50 1 50 99 2 100 299 3 300 499 4 500 999 5 6 EN % < 5 1 5 10 2 10 25 3 25 50 4 50 75 5 75 100 6 EN HAS < 50 1 50 99 2 100 299 3 300 49 9 4 500 999 5 6 NO CONTESTA 9

PAGE 135

135 Q3 3 Podra mencionarme qu cultivos sembr esta ltima campaa (20067 PERODO DE REFERENCIA: 1 AGOSTO 2006 31 MAYO 2007) y cunta superficie de los mismos (en has o en % de superficie total de la explotacin)? EN HAS EN % ID CULTIVO 1 MAZ 2 GIRASOL 3 SOJA 4 TRIGO 5 SORGO 6 ALFALFA 7 MOHA 8 MIJO 9 OTRO (ESPECIFICAR) ___________________ NO CONTESTA 9 Q3 4 Sera tan amable de indicarnos, por favor, el porcentaje aproximados de sus ingresos anuales que proviene de la agricultura? Q3 5 Realiza otras actividades productivas, si o no (ganadera, avicultura, tambo, etc.) ? NO.. 1 IR A Q 38 SI........ 2 NO CONTESTA. 9 Q3 6 Podra indicarme, a continuacin, cuntas hectreas (o porcentaje aproximado del total) destin, en esta ltima campaa ( PERODO DE REFERENCIA: 1 AGOSTO 2006 31 MAYO 2007), a estas otras actividades productivas? ______________________________ has NO CONTESTA 9 EN % < 5 1 5 10 2 10 25 3 25 50 4 50 75 5 75 100 6 EN HAS < 50 1 50 99 2 100 299 3 300 499 4 500 999 5 6 EN % < 5 1 5 10 2 10 25 3 25 50 4 50 75 5 75 100 6 NO CONTESTA 9

PAGE 136

136 Q 37 Sera tan amable de indicarnos, por favor, el porcentaje aproximados de sus in gresos anuales que proviene de estas otras actividades productivas? Q 38 Si le parece, hablemos ahora de montes en su explotacin Tiene sitios con monte en su explotaci n, si o no ? NO.. 1 IR A Q4 2 SI........ 2 NO CONTESTA. 9 Q 39 Podra indicarme si los sitios corresponden a monte nativo, introducido (distinto que eucaliptus), eucaliptus, o mixto? Q4 0 Cunto ocuparan, en hectreas o porcentaje aproximado, los sitios de monte de la superficie total de su explotacin ? ______________________________ has Q4 1 Cul de las siguientes opciones es la ms probable respecto a los sitios de monte que tiene actualmente: que mantenga los sitios tal como estn, que aumente los siti os con monte, o desmonte (parte o todo)? ID ACCION MONTE NATIVO MONTE INTRODUCIDO MONTE MIXTO 1 DESMONTAR TODO 1 2 3 2 MANTENER LOS SITIOS COMO ESTAN 1 2 3 3 AUMENTAR 1 2 3 4 NO CONTESTA 9 Q4 2 A continuacin, y ya para ir cerrando la entrevista, vamos a realizar algunas preguntas sobre su participacin en actividades sociales relacionadas con el agro. Por favor, podra indicarnos si pertenece o no a una organizacin de productores o relacionada con la agricultura? NO.. 1 IR A Q4 4 S I........ 2 NO CONTESTA. 9 EN % < 5 1 5 10 2 10 25 3 25 50 4 50 75 5 75 100 6 NO CONTESTA 9 TIPO DE MONTE NO SI 1 NATIVO 1 2 2 DE EUCALIPTUS 1 2 3 INTRODUCIDO DISTINTO DE EUCALIPTUS 1 2 4 MIXTO CON EUCALIPTUS 1 2 5 MIXTO SIN EUCALIPTUS 1 2 NO CONTESTA 9 EN % < 5 1 5 10 2 10 25 3 25 50 4 50 75 5 75 100 6 EN HAS < 50 1 50 99 2 100 299 3 300 499 4 500 999 5 6 NO CONTESTA 9

PAGE 137

137 Q4 3 A cuntas organizaciones? 1 1 2 3 2 3 5 3 >5 4 NO CONTESTA 9 Q4 4 Participa usualmente en jornadas o reuniones agrcolas? NO.. 1 IR A Q 46 SI........ 2 NO CONTESTA. 9 Q 45 Cuntas reuniones al ao ? Q 46 Pertenece a una organizacin ambientalista o relacionada con la conservacin de la naturaleza? NO.. 1 IR A Q 49 SI........ 2 NO CONTESTA. 9 Q 47 Cuntas organizaciones? Q48 Podra nombrarla/s, por favor? Q 49 Cules son sus fuentes primarias de informacin sobre manejo de plagas, incluyendo manejo de conflicto s con aves?. Si lo desea, puedo leerle un listado de posibles fuentes de informacin. All puede indicar ms de una fuente. NO SI 1 Universidad 1 2 2 Extensin de una Cooperativa 1 2 3 Agentes del gobierno nacional 1 2 4 Agentes del gobierno provin cial 1 2 5 Agroqumicas 1 2 6 Otros productores 1 2 7 Medios masivos de comunicacin 1 2 8 Otros (especificar) _____________________________ 1 2 9 NO CONTESTA 1 2 < 6 (1 cada 2 meses o menos) 1 6 12 (entre 1 c/2 meses y 1 por mes) 2 > 12 (ms de 1 por mes) 3 NO CONTESTA 9 1 1 2 3 2 3 5 3 >5 4 NO CONTESTA 9

PAGE 138

138 APPENDIX D MOSAIC GRAPHICS FOR RELATIONSHIPS AMONG INDEPENDENT VARIABLES -1 0 2 4 Pearson residuals: p-value = 3.072e-06 Education Age Attitude 50-59 yr Po Ne Mo 40-49 yr Po Ne Mo > 60 yr Po Ne Mo < 40 yr IPS PS SS UI Po Ne Mo Figure D 1. Mosaic display showing the relationships among age, education and attitudes toward monk parakeets T he area of each rectangle is proportional to the observed frequency of farmers in that rectangle (Friendly 2000, pg. 106). Colors indicate deviations from independence, in this case, higher frequency than would be f ound under independence (in dark grey ). Education: IPS= incomplete primary school, PS= primary school, SS=secondary school, UI=university instruction. Attitudes: PO= positive, NE= negative, MO= moderate.

PAGE 139

139 -1.48 0.00 2.00 2.14 Pearson residuals: p-value = 0.064266 attitude effecttot RC MIP LC HM CR CP AP MO NE PO Figure D 2. Relationships among attitudes and beliefs about effectiveness of management strategies. Attitudes: PO= positive, NE= negative, MO= moderate. Beliefs about the most effective management strategy: LC= Lethal Control, LC= lethal con trol, CP= crop protection, AP= agricultural practices, HM= habitat management, RC= reproductive control, CR= capture and release, IPM= integrated pest management.

PAGE 140

140 -1 0 1 Pearson residuals: p-value = 0.13417 subjnorm effecttot RC MIP LC HM CR CP AP HI LO MO Figure D 3. Relationships between beliefs about the most e ffective management strategy and influence of subjective norms. Management strategy: LC= Lethal Control, LC= lethal control, CP= crop protection, AP= agricultural practices, HM= habitat management, RC= reproductive control, CR= capture and release, IPM= integrated pest management. Inf luence of subjective norms: LO= low, MO=moderate, HI=high.

PAGE 141

141 LIST OF REFERENCES Addicott, F., Aho, J.M. Antoli, M.F. Padilla, D.K., Richardson J.S. & Soluk D.A. (1987) Ecological neighborhoods: scaling environmental patterns. Oikos 49 340 346. Agrest i, A. (2002) Categorical data analysis. Willey Series in Probability and Statistics. 2nd edn. Wiley Interscience. Ajzen, I. (1985) From intentions to actions: A theory of planned behavior. Action control: From cognition to behavior (eds J. Kuhl & J. Beckma nn), pp. 1139. Springer Verlag, Berlin, Germany. Ajzen, I. (1991) The theory of planned behavior. Organizational Behavior and Human Decision Processes 50 179 211. Amano, T., Ushiyama, K. & Higuchi, H. (2008) Methods for predicting risks of wheat damage by whitefronted geese. Journal of Wildlife Management 72, 18451852. Amano, T., Ushiyama, K., Fujita G. & Higuchi, H. (2004) Alleviating grazing damage by white fronted geese: an optimal foraging approach. Journal of Applied Ecology 41, 675688. Am ano, T., Ushiyama, K., Fujita G. & Higuchi, H. (2007) Predicting grazing damage by white fronted geese under different regimes of agricultural management and the physiological consequences for the geese. Journal of Applied Ecology 44, 506 515. Arambur, R .M (1991) Contribucin al estudio biolgico de la cotorra Myiopsitta monachus en la provincia de Buenos Aires (Aves: Psittacidae) PhD thesis, University of La Plata, Argentina. Arambur RM (1997) Ecologa alimentaria de la cotorra Myiopsitta monachus mona chus en la provincia de Buenos Aires, Argentina (Aves: Psittacidae). Physis, 53, 29 32 Arambur RM (1998) La cotorra ( Myiopsitta monachus ) en la provincia de Buenos Aires. Manual de Capacitacin sobre Manejo Integrado de Aves Perjudiciales a la Agricultur a (eds E. Rodrguez & M.E. Zaccagnini), pp. 5052. United Nations, Food and Agriculture Organization (FAO). Binacional Project ArgentinaUruguay. TCP/RLA/8965. 171 pp. Arambur, R.M. & Bucher, E.H. (1999) Preferencias alimentarias de la Cotorra Comn Myiop sitta monachus (Aves: Psittacidae) en cautividad. Ecologa Austral 9 11 14. Avery, M.L. (2002) Birds in Pest Management. Encyclopedia of Pest Management (ed. D. Pimental), pp.104106. Marcel Dekker, New York.

PAGE 142

142 Avery, M.L. Greiner, E.C., Lindsay, J.R., N ewman, J.R., & PruettJones, S. (2002) Monk parakeet management at electric utility facilities in South Florida. Proceedings 20th Vertebrate Pest Conference (eds R.M. Timm and R.H. Schmidt), pp.1401 45. University of California, Davis Avery, M. L., Yode r, C. A. & Tillman E. A. (2008) Diazacon inhibits reproduction in invasive monk parakeet populations. Journal of Wildlife Management 72 14491452. Balley, J.W., & Thompson, F.R. (2007) Multiscale Nest Site Selection by Black Capped Vireos. The Journal of Wildlife Management 71, 828836. Batry, P., Bldi, A., Kleijn, D. & Tscharntke, T. (2011) Landscapemoderated biodiversity effects of agri environmental management: a metaanalysis. Proceedings of the Royal Society B 278 1894 1902. Beletsky, L.D. ( 1996) The redwinged blackbird: the biology of a strongly polygynous songbird. Academic Press, San Diego, CA. Best, L.B., Bergin, T.M. & Freemark, K.E. (2001) Influence of Landscape Composition on Bird Use of Rowcrop Fields. Journal of Wildlife Management 65, 442449. Bianchi, F.J.J.A., Booij, C.J.H. & Tscharntke, T. (2006) Sustainable pest regulation in agricultural landscapes: a review on landscape composition, biodiversity and natural pest control. Proceedings of the Royal Society B. 273 1715 1727. Bi bby, C.J., Burgess, N.D., Hill, D.A. & Mustoe, S.H. (2000). Bird census techniques 2nd edn. Academic Press, London. 302 pp. Bjerke, T., Reitan O. & Kellert, S.R. (1998) Attitudes toward wolves in southeastern Norway. Society and Natural Resources 11, 169178. Bomford, M. & Sinclair, R. (2002) Australian research on bird pests: impact, management y future directions. Emu 102 29 45. Boscolo, D. & Metzger, J.P. (2009) Is bird i ncidence in Atlantic forest fragments Landscape Ecology 24 907 918. Boutin, C., Freemark, K.E. & Kirk, D.A. (1999a) Spatial and Temporal Patterns of Bird Use of Farmland in Southern Ontario. Canadian Fi eld Naturalist 113 430 460. Boutin, C., Freemark, K.E. & Kirk, D.A. (1999b) Farmland Birds in Southern Ontario: Field Use, Activity Patterns and Vulnerability to Pesticide Use. Agriculture, Ecosystems and Environment, 72 239 254.

PAGE 143

143 Brennan, J. M., Bender, D. J. Contreras, T. A., & Fahrig L. ( 2002) Focal patch landscape studies for wildlife management: optimizing sampling effort across scales. Integrating Landscape Ecology into Natural Resource Management. (eds J. Liu & W. W. Taylor ), pp. 6891. Cambrid ge University Press, Cambridge. Brennan, S.P. & Schnell G.D. (2005) Relationship between bird abundances and landscape characteristics: the influence of scale. Environmental Monitoring and Assessment 105 209228. Brown, T.C. & Peterson, G.L. (2003) Multi ple Good Valuation, With Focus on the Method of Paired Comparisons. A Primer on Nonmarket Valuation (eds P. Champ, K.J. Boyle & T.C. Brown), pp. 221258. Kluwer Academic Publishers. Brown, T.L., Decker, D.J. & Dawson, C.P. (1978) Willingness of New York f armers to incur whitetailed deer damage. Wildlife Society Bulletin 6 235 239. Bruggers, R.L. & Elliott, C.C.H. (1989) Quelea quelea: Africa's bird pest. Oxford University Press. 402 pp. Bruggers, R.L. & Zaccagnini, M.E. ( 1994) Vertebrate pest problems related to agricultural production and applied research in Argentina. Vida Silvestre Neotropical, 3 71 83. Bruggers, R.L., Owens, R. & Hoffman, T. (2002) Wildlife damage management research needs: perceptions of scientists, wildlife managers, and stakeholders of the USDA/Wildlife Services program. International Biodeterioration & Biodegradation, 49, 213 223. Bruggers, R.L., Rodriguez, E. & Zaccagnini, M.E. (1998) Planning for bird pest problem resolution: a case study. International Biodeterioration and Bi odegradation, 42 173184. Bruun, M. & Smith H.G. ( 2003) Landscape composition affects habitat use and foraging flight distances in breeding European starlings. Biological Conservation, 114 179 187. Bucher, E.H. & Bedano, P.E. (1976) Bird Damage Problems in Argentina. International Studies on Sparrows 9 3 16. Bucher, E.H. (1984) Las aves como plaga en Argentina. 9th edn. Centro de Zoologa Aplicada. Crdoba, Argentina. Bucher, E.H. (1985) Ecologa de aves plaga en el Uruguay. Consultant report. United Nations, Food and Agricultura Organization (FAO) Binational Project. ArgentinaUruguay. TCP/RLA/8965. 15 pp.

PAGE 144

144 Bucher E .H. ( 1990) The influence of changes in regional landuse patterns on zenaida dove populations. Granivorous Birds in Agricultural Landscape (eds Pinowsky J. & Summers Smith J. D .). pp. 291 303 Polish Academy of Sciences Warshaw. Bucher, E.H. (1992a) Neotropical Parrots as Agricultural Pests. New World Parrots in Crisis. Solutions from Conservation Biology (eds S. Beissinger & N.F. Snyder), pp. 201219. Smithsonian Institution Press. Bucher, E.H. (1992b) Aves Plaga de Argentina y Uruguay: Dinmica de poblaciones. Consultant Report. United Nations, Food and Agriculture Organization (FAO). Binational Project ArgentinaUruguay. TCP/RLA/8965. 18 pp. Bucher, E.H. (1998) Criterios bsicos para el Manejo Integrado de Aves Plaga. Manual de Capacitacin sobre Manejo Integrado de Aves Perjudiciales a la Agricultura (eds E. Rodrguez & M.E Zaccagnini), pp.6170. United Nations, Food and Agriculture Orga nization. Binational project. ArgentinaUruguay. 171 pp. Bucher, E.H. & Martin, L.F. (1987) Los nidos de cotorras ( Myiopsitta monachus ) como causa de problemas en lneas de trasmisin elctrica. Vida Silvestre Neotropical 1 50 51. Bucher, E.H., Martin, L .F., Martella, M. & Navarro, J. (1991) Social behavior and population dynamics of the Monk Parakeet. Acta XX Congressus Internationalis Ornithologici 2 681 689. Christchurch, New Zealand. Bucher, E.H. & Ranvaud, R.D. (2006) Eared dove outbreaks in South America: patterns and characteristics. Acta Zoologica Sinica, 52 564 567 Burel, F. & Baudry, J. (1995) Farming Landscapes and Insects. Ecology and Integrated Farming Systems. (eds D.M. Glen, M.P. Graves & H.M. Anderson), pp. 204220. John Wiley and Sons Ltd., Chichester, England. 329 pp. Burger, J. & Gochfeld, M. (2000) Nest site selection in monk parakeets ( Myiopsitta monachus ) in Florida. Bird Behavior 13 99 105. Burger, J. & Gochfeld, M. (2005) Nesting behavior and nest site selection in monk parake ets ( Myiopsitta monachus ) in the Pantanal of Brazil. Acta Ethologica, 8 23 34. Burger, J. & Gochfeld, M. (2009) Exotic monk parakeets ( Myiopsitta monachus ) in New Jersey: nest site selection, rebuilding following removal, and their urban wildlie appeal. U rban Ecosyst em 12, 185 196. Burgess, D., Hutchinson, W. G., McCallion, T. & Scarpa R. (2002). Investigating Choice Rationality in Stated Preference Methods for Enhanced Farm Animal Welfare. The Centre for Social and Economic Research on the Global Environment (CSERGE) Working Paper. ECM 200302. University of East Anglia, United Kingdom.

PAGE 145

145 Burham, K.P & D.R. Anderson. (2002) Model Selection and Multimodel Inference. 2nd edn. Springer Verlag, New York, USA. Burghardt, K.T., Tallamy, D.W., & Shriver W.G. (20 08) Impact of n ative p lants on b ird and b b iodiversity in suburban l andscapes. Conservation Biology 23, 219 224. Butterfield, R.P. (1995) Promoting biodiversity: advances in evaluating native species f or r eforestation. Forest Ecology and Management 75, 111121. Campa, H., Winterstein, S.R., Peyton, R.B., Dudderar, G.R., Leefers, L.A. (1997) An evaluation of a multidisciplinary problem: ecological and sociological factors influencing whitetailed deer damage to agricultural crops in Michigan. Transactions of the. North American Wildlife and Natural Resource Conference, 62, 431440. Canavelli, S.B. & Arambur, R.M. ( in press ). Aspectos a considerar en el manejo de conflictos entre las cotorras ( Myiopsitta monachus ) y la produccin agrcola. El Hornero, n.a.. Canavelli, S.B. & Zaccagnini, M.E. (2007) Nuevos enfoques en el manejo de conflictos con fauna silvestre para una agricultura sustentable. Agricultura Sustentable en Entre Ros (eds O.P. Caviglia, O.F. Paparotti & M.C. Sasal), pp 205214. Ediciones INTA. Buenos Aires. Canavelli, S.B. (2003) Projected impacts of lethal control on monk parakeet populations in Argentina. Final paper. Wildlife Population Modeling course (WIS 6466). University of Florida, Gainesville. Canavelli, S.B., Gonzlez, C., Cav allero, P. & Zaccagnini, M.E. (2008) Dao relativo por aves en cultivos de maz y girasol del departamento Paran y zonas aledaas. Serie de Extensin de la EEA Paran, 51, 59 67. Cavallero, P.L. (2010) Nidificacin de la cotorra comn en montes nativos: anlisis a escala de parche y de paisaje. Tesina de Grado. Universidad Nacional del Litoral. Santa Fe, Argentina. 47 pp. Clark, R.G., Titman, R.D., Roger Bider, J. & Prvost, M. (1982) Aspects of the socio economic impact of blackbirds and other vertebrates on farm production in Qubec. Phytoprotection, 63, 56 67. Clergeau, P.H. (1995) Importance of Multiple Scale Analysis for Understanding Distribution and for Management of an Agricultural Bird Pest. Landscape and Urban Planning, 31 281 289. Cochran, W.G. (1977) Sampling Techniques 3rd edn. John Wiley and Sons. New York, USA.

PAGE 146

146 Cody, M.L. (1985) Habitat selection in birds Aademic Press, Orlando, Florida, USA. Colbach, N., Dessaint, F. & Forcella, F. (2000) Evaluating fieldscale sampling methods for the es timation of mean plant densities of weeds. Weed Research, 40 411 430. Colluccy J. M., Drobney, R D. Graber, D A. Sheriff, S L. & Witter D J. (2001) Attitudes of central Missouri residents toward local giant Canada geese and management alternatives W ildlife Society Bulletin, 29 ,116 123. Conover, M. (2002) Resolving HumanWildlife Conflicts:The Science of Wildlife Damage Management 1st edn. Lewis Publisher. Florida, USA. 418 pp. Cooper, C.B. & Walters, J.R. (2002) Independent effects of woodland loss and fragmentation on Brown Treecreeper distribution. Biological Conservation, 105 1 10. De Grazio, J. W. (1985) Bird Pest Problems in Uruguay with special notations on damage appraisal methodology. Consultant Report. United Nations, Food and Agriculture O rganization (FAO). Binational Project ArgentinaUruguay. TCP/URU/4401.18 pp. De Grazio, J.W. & Besser, J.F. (1975) Field investigations of monk parakeet damage to sunflowers and corn in Uruguay. Denver Wildlife Research Center. Trip Report. 23 pp. De Grazi o, J.W. (1978) World bird damage problems. Proceedings of 8th Vertebrate Pest Conference. Univers i ty of Nebraska, Lincoln, 8 9 24. De Grazio, J.W., Besser, J.F., Guarino, J.L., Loveless, C.M. & Oldemeyer, J.L. (1969) A method for appraising blackbird damage to corn. Journal of Wildlife Management 33, 988994. de Vaus, D.A. 2001. Research Design in Social Research SAGE Publications. London, UK. 279 pp. Decker, D.J. & Brown, T.L. (1982) Fruit growes vs. other farmer attitudes toward deer in New York Wildlife Society Bulletin, 10 150155. del Hoyo, J. Elliot, A. & Sargatal, J. (1992) Handbook of the birds of the world. Sandgrous to Cucos. Lynx Ediciones. Barcelona. DeVellis, R.F. (2003) Scale development: theory and applications, 2nd edn. Applied Social Research Methods Series, Sage Publications. Di Rienzo, J.A., Casanoves, F., Balzarini, M.G., Gonzalez, L., Tablada, M., Robledo, C.W. (2010) InfoStat versin 2010. Grupo InfoStat, Universidad Nacional de Crdoba, Facultad de Ciencias Agropecuarias, Argent ina.

PAGE 147

147 Dolbeer, R.A. (1975) A comparison of two methods for estimating bird damage to sunflowers. Journal of Wildlife Management 39, 802806. Domenech, J., Carrillo, J. and Senar, J.C. (2003) Population size of the monk parakeet Myiopsitta monachus in Catal onia. Revista Catalana dOrnitologia, 20, 1 9. Donzar J.A., Hiraldo, F., & Bustamante, J. (1993) Factors influencing nest site selection, breeding density and breeding success in the bearded vulture ( Gypaetus barbatus ). Journal of Applied Ecology 30, 50 4 514. Don Carlos, A.W., Bright, A.D., Teel, T.L. & Vaske J.J. (2009). Human black bear conflict in urban areas: an integrated approach to management response. Human Dimensions of Wildlife, 14, 174 184. Duke, J.A. (1983) Handbook of Energy Crops. Purdue University, Center for New Crops and Plant Products. U npublished. Available from : http://www.hort.purdue.edu/newcrop/indices/index_efgh.html#E (Last day accessed: July 22, 2011 ). D unning, J.B., Danielson, B.J. & Ronald Pulliam, H. (1992) Ecological Processes That Affect Populations in Complex Landscapes. Oikos 65 169 175. Dyer, M.I. & Ward, P. (1977) Management of pest situations. Granivorous birds in ecosystems (eds J. Pinowsky & S.C. Kendeigh), pp. 267300. Cambridge University Press, Cambridge. Eberhard, J.R. (1998) Breeding biology of the monk parakeet. Wilson Bulletin 110 463 473. Engler, P.L. & Vicente, G.R. (2009) Caracterizacin y localizacin de las formas de organizaci n social predominantes de las unidades agropecuarias de Entre Ros. Proceedings of the X Congreso Argentino de Ingeniera Rural y II del MERCOSUR. Rosario, Argentina. Fahrig, L. (2001) How much habitat is enough? Biological Conservation, 100 65 74. Fa l l M.W. & Jackson, W.B. (200 0 ) Future technology for managing problems with vertebrate pests and over abundant wildlifean introduction. International Biodeterioration & Biodegradation 45 93 95. Fargestone, K .A., Miller, L.A., Killian, G. & Yoder, C.A. (2010) Review of issues concerning the use of reproductive inhibitors, with particular emphasis on resolving humanwildlife conflicts in North America. Integrative Zoology 1 1530. Feare, C.J. (1993) Control of bird pest populations. Bird Population Studies (eds C.M Perrins, J.D. Lebreton & G.J.M. Hirons), pp. 463478. Oxford University Press.

PAGE 148

148 Fishbein, M. & Ajzen, I. (1975) Belief, Attitude, Intention and Behaviour: An Introduction to Theory and Research. Reading, MA: AddisonWesley. Out of Print. Availa ble from: http://people.umass.edu/aizen/f&a1975.html (Last day accessed: July 22, 2011). Fletcher, R.J. Jr & Koford, R.R. (2002) Habitat and landscape associations of breeding birds in native and restored grasslands Journal of Wildlife Management 66, 10111022. Forman, R.T.T. and M. Godron. (1986) Landscape Ecology 1st edn. John Wiley and Sons. New York, USA. 619 pp. Forshaw, J.M. (1989) Parrots of the world. 3rd edn. Landsdowne Editions, Melbourne. Australia. Freeland D.B. (1973) Some food preference and aggressive behavior by Monk Parakeets. Wilson Bulletin 85, 332 333. Freemark, K. & Rogers, C. (1995) Modification of point counts for surveying cropland birds. Monitoring Bird Populations by P oint Counts (eds C.J. Ralph, J.R. Sauer & S. Droege), pp. 6974. Albany, CA. Friendly, M. (2000) Visualizing Categorical Data. Cary NC: SAS Institute Inc. 435 pp. Fritzell, P.A., Minnis, D.L. & Peyton, R.B. (1997) A comparison of deer hunter and farmer att itudes about crop damage abatement in Michigan: messages for hunters, farmers and managers. Proceedings of the Eastern Wildlife Damage Management Conference, 7 153 161. Gimnez, J.I. & Salomn, C. (1999) Evaluacin del dao provocado por las aves plaga en los cultivos de girasol. Unpublished report. Gobierno de la Provincia de Entre Ros. Secretara de la Produccin. 16 pp. Gimnez, J.I. and C.Salomn (2000) Anlisis de la problemtica de las aves plagas en distintas zonas de la provincial Unpublished report. Gobierno de la Provincia de Entre Ros. Secretara de la Produccin. 10 pp. Haggara, J.P., Briscoe, C.B. & Butterfield, R.P. (1998) Native species: a resource for the diversification of forestry production in the lowland humid tropics. Forest Ecology and Management 106 195 203. Hagy, H.M., Linz, G.M & Bleier, W.J. (2008) Optimizing the use of decoy plots for blackbird control in commercial sunflower. Crop Protection, 27, 14421447. Heong, K.L. & Escalada, M.M. (1999) Quantifying rice farmers pest management decisions: belief and subjective norms in stem borer control. Crop Protection, 18, 315322.

PAGE 149

149 Heong K. L., Escalada, M. M., Sengsoulivong, V. & Schiller J. (2002) Insect management beliefs and practices of rice farmers in Laos Agriculture, Ecosyste ms and Environment 92 137 145. Heong, K. L., Escalada, M. M. & Vo Mai (1994) An analysis of insecticide use in rice: case studies in the Philippines and Vietnam. International Journal of Pest Management 40 ,173 178. Hildn, O. (1965) Habitat selection in birds: a review. Annales Zoologici Fennici 2 53 75. Hines, J., Hungerford, H. & Tamera, A. ( 1986) Analysis and synthesis of research on environmental behavior: a meta analysis. Journal of Environmental Education, 18 1 8. Holland, J.D., Bert D.G. & Fah rig L ( 2004) Determining the Spatial Scale of Species Response to Habitat. BioScience, 54 227233. Holling, C.S. 1978. Adaptive environmental assessment and management. John Wiley & Sons. London. Holt, R.D., Debinski D.M., Diffendorfer J.E., Gaines M.S ., Martinko E.A., Robinson G.R. and Ward G.C. (1995) Perspectives from and Experimental Study of Habitat Fragmentation in an Agroecosystem. Ecology and Integrated Farming Systems. (eds D.M. Glen, M.P. Graves & H.M. Anderson), pp. 147175. John Wiley and S ons Ltd., Chichester, England. 329 pp. Hone, J. (1994) Analysis of vertebrate pest control Cambridge University Press, New York, USA. 258 pp. Hrubes, D., Ajzen, I. & Daigle, J. (2001) Predicting hunting intentions and behavior: an application of the Theor y of Planned Behavior. Leisure Sciences 23, 165178. Humphrey, P.S. and Peterson R.T. (1978) Nesting behavior and affinities of the monk parakeets of southern Buenos Aires province, Argentina. Wilson Bulletin 90 5 445 52. Hyman, J., & S. Pruett Jones (1 995) Natural history of the monk parakeet in Hyde Park, Chicago. Wilson Bulletin 107 510517. Instituto Nacional de Estadsticas y Censos ( 2002) Censo Nacional Agropecuario (CNA). Argentina. Jacobson S.K., Sieving, K.E., Jones, G.A. and Van Doorn, A. (2003) Assessment of Farmer Attitudes and Behavioral Intentions toward Bird Conservation on Organic and Conventional Florida Farms. Conservation Biology, 17 595606.

PAGE 150

150 Johnson, D. H. (1980) The comparison of usage and availability measurements for evaluating resource preference. Ecology 20 553 562. Jonker, S.A., Muth, R.M., Organ, J.F., Zwick, R.R., & Siemer, W.F. (2006) Experiences with beaver damage and attitudes of Massachusetts residents toward beaver. Wildlife Society Bulletin 34 1009 1021. Kahn, H.A. (2003) Movement models of the Roseringed Parakeet ( Psittacula krameri ) in Daylight Hours in its Communal Roost. Pakistan Journal of Biological Sciences 6 184 187. Keith, J.O. (1991) Bird Pests in Argentina and Uruguay: Ecotoxicological Evaluation of Co ntrol Programs Consultant Report. United Nations, Food and Agriculture Organization (FAO).TCP/RLA/8965(A). Kellert, S.R. & Berry, J.K. (1987) Attitudes, knowledge, and behaviors toward wildlife as affected by gender. Wildlife Society Bulletin 15 363 371 Kellert, S.R. (1980) Americans attitudes and knowledge of animals. Transactions of the North American Wildlife and Natural Resources Conference, 45 111 124. Kogan, M. (1998) Integrated pest management: historical perspectives and contemporary developments. Annual Review of Entomology 43, 243 270. Koper, N. & Schmiegelow, F.K.A. (2006) A multi scaled analysis of avian response to habitat amount and fragmentation in the Canadian dry mixedgrass prairie. Landscape Ecology 21 1045 1059. Koval, M.H. & Mer tig, A.G. (2004) Attitudes of the Michigan public and wildlife agency personnel toward lethal wildlife management. Wildlife Society Bulletin 32 232 243. Landis, D.A. and Marino P.C. (1999) Landscape Structure and ExtraField Processes: Impact on Management of Pests and Beneficials. Handbook of Pest Management. (ed. Ruberson, J.R.), pp. 79104. Marcel Dekker, Ink. New York, USA. Li, P. & Martin, T.E. ( 1991) Nest site selection and nesting success of cavity nesting birds in high elev ation forest drainages. Auk 108, 405 418. Linz, G.M. (1999) Assessing Blackbird Damage to Sunflower Fields National Wildlife Research Center. Standard Operating Procedure. FP/ND 006.00. Linz, G.M., Peer, B.D., Homan, H.J., Wimberly, R.L., Bergman, D.L., Bleier, W.J. (2002) Has an integrated pest management approach reduced blackbird damage to sunflower? Human Conflicts With Wildlife: Economic Considerations. (eds Clark, L., Hone, J., Shivik, J.S., Watkins, R.W., VerCauteren, K.C., Yoder, J.K.), pp. 132 137. Proceedings of the 3r d NWRC Special Symposium of the National Wildlife Research Center, Fort Collins, Colorado, USA.

PAGE 151

151 Loker, C.A., Decker, D.J. & Schwager, S.J. (1999) Social acceptability of wildlife management actions in suburban areas: 3 cases from New York. Wildlife Society Bulletin 27,152 159. Loyd, K.A.T. & Miller, C.A. (2010) Influence of demographic, experience and value orientations on preferences for lethal management of feral cats. Human Dimensions of Wildlife 15, 262 273. Martin, L. F. and Bucher, E.H. (1993) Natal dispersal and first breeding age in Monk Parakeets. Auk 110 930 933. Martin, S.R. & McCurdy, K. (2009) Wilderness food storage in Yosemite: using the Theory of Planned Behavior to understand backpacker canister use. Human Dimensions of Wildlife 14, 206 218. Martinez J.A., Serrano, D. & Zuberogoitia, I. (2003) Predictive models of habitat preference for the Eurasian eagle owl Bubo bubo: a multiscale approach. Ecography 26, 21 28. Mazerolle, M.J. & Villard, M. (1999) Patch characteristics and landscape c ontext as predictors of species presence and abundance: a review. Ecoscience 6 117 124. McGarigal, K., Cushman, S.A., Neel, M.C. & Ene E (2002) FRAGSTATS: Spatial Pattern Analysis Program for Categorical Maps. Computer software program produced by the authors at the University of Massachusetts, Amherst. Available from : http://www.umass.edu/landeco/research/fragstats/fragstats.html (Last day accessed: July 22, 2011). Messmer, T .A. ( 2000) Emergence of human wildlife conflict management: turning challenges into opportunities. International Biodeterioration 45 97 100. Messmer, T.A. & Schroeder, S. (1996) Perceptions of Utah alfalfa growers about wildlife damage to their hay crops: implications for managing wildlife on private land. Great Basin Naturalist 56, 254 260. Messmer, T.A., Cornicelli, L., Decker, D.J. and Hewitt, D.G. (1997) Stakeholder acceptance of urban deer management techniques. Wildlife Society Bulletin 25 3603 66. Moschione, F.N. & Banchs, R.A. (2006) Proyecto Calas: U na experiencia de manejo adaptativo para el aprovechamiento sustentable de psitcidos y como estrategia de conservacin de sus hbitats en la Argentina. Manejo de Fauna Silvestre en la Argentina. P rogramas de uso sustentable (eds M.L. Bolkovic & D. Ramadori), pp. 2737. Secretara de Ambiente y Desarrollo Sustentable, Buenos Aires. Argentina. Mumford, J.D. & Norton, G.A. (1984) Economics of Decision Making in Pest Management. Annual Review of Entomology 29, 157 174.

PAGE 152

152 Navarro J.L., Martella, M.B. and Bucher, E.H. (1992) Breeding Season and Productivity of Monk Parakeets in Cordoba, Argentina. Wilson Bulletin 104 413424. Norton, G.A. & Mumford J.D. (1993) Decision Tools for Pest Management CAB In ternational. University Press, Cambridge, UK. 279 pp. Orians, G. (1985) Blackbirds of the Americas University of Washington Press. Otis, D.L. (1992) Recommendations on the Need and Use of Statistical Surveys for Evaluating the Impacts of Bird Damage to Cr ops in Uruguay and Argentina. Consultant report. United Nations, Food and Agriculture Organization (FAO). Binational Project ArgentinaUruguay. TCP/RLA/8965. 22 pp. Otis, D.L. & Kilburn, C.M. (1988) Influence of Environmental Factors on Blackbird Damage to Sunflower. United States Department of Interior, Fish and Wildlife Service Technical Report 16. Washington D.C. Parkes J.P., Robley A., Forsyth D.M. & Choquenot D. (2006) Adaptive Management Experiments in Vertebrate Pest Control in New Zealand and Australia. Wildlife Society Bulletin 34 229 236. Pierce, C.L., Manfredo, M.J. & Vaske, J.J. (2001) Social Science Theories in Wildlife Management. Human dimensions of wildlife in North America (eds D.J. Decker, D.J., T.L. Brown & W.F. Siemer), pp. 39 56. The W ildlife Society. Pinowski. J. & Kendeigh S.Ch. (1977) Granivorous birds in ecosystems. Cambridge University Press, UK. Power, A.G. (2010) Ecosystem services and agriculture: tradeoffs and synergies. Philosophical Transactions of the Royal Society B 365 2 959 2971 Reiter, D.K., Brunson, M.W. & Schmidt, R.H. (1999) Public attitudes toward wildlife damage management and policy. Wildlife Society Bulletin 27 746758. Rempel. R. (2010) Patch Analyst extension for ArcView GIS Available from: http://flash.lakeheadu.ca/~rrempel/patch/ (Last day accessed: July 22, 2011). Renfrew, R.B. & Ribic, C.A. (2008) Multiscale models of grassland passerine abundance in a fragmented system in Wisconsin. Landscape Ec ology 23, 181 193. Rossi, A.N. & Armstrong, J.B. (1999) Theory of Reasoned Action vs. Theory of Planned Behavior: testing the suitability and sufficiency of a popular behavior model using hunting intentions. Human Dimensions of Wildlife, 4 40 56.

PAGE 153

153 Santos, D.M. (2005) Myiopsitta monachus. En: Fichas de aves introducidas en Espaa. Grupo de Aves Exticas (SEO/BirdLife). Available from : http://www.seo.org/media/docs/f_myiopsitta_monachus .html (Last day accessed: July 22, 2011). Savary, S. (1993) Rice farmers background, perceptions of pests, and pest management actions: a case study in the Philippines. Netherlands Journal of Plant Pathology 99 181. Shea K., Possingham, H.P., Murdoch, W.W. & Roush, R. (2002) Active adaptive management in insect pest and weed control: intervention with a plan for learning. Ecological Applications 12, 927 936. Soh, M.C.K., Sodhi, N.S., Seoh, R.K.H. and Brook, B.W. (2002) Nest site selection of the house crow ( Corvus splendens ), an urban invasive bird species in Singapore and implications for its management. Landscape and Urban Planning 59, 217 226. Sol, D., Santos, D.M., Feria, E. & Clavell, J. (1997) Habitat selection by the monk parakeet during coloniz ation of a new a rea in Spain Condor 99, 39 46. Spreyer, M.F. & Bucher, E.H. (1998) Monk parakeet ( Myiopsitta monachus ). The Birds of North America (eds A. Poole & F. Gill), 322, 1 23. Philadelphia, PA. Stephens, D. W. & J.R. Krebs (1986) Foraging theory. Princeton University Press, Princeton, NJ. Stout, R.J., Knuth, S.A. & Curtis, P.D. (1997) Preferences of suburban landowners for deer management techniques: a step towards better communication. Wildlife Society Bulletin 25 348 359. Stout, R.J., Stedman, R.C., Decker, D.J. & Knuth, B.A. (1993) Perceptions of risk from deer related vehicle accidents: implications for public preferences for deer herd size. Wildlife Society Bulletin, 21, 237249. Surmacki, A. (2005) Habitat use by three Acrocephalus warblers in an intensively used farmland area: the influence of breeding patch and its surroundings. Journal of Ornithology 146 160 166. Thornton, D.H., Branch, L.C. & Sunquist, M.E. (2010) The influence of landscape, patch, and withinpatch factors on species pr esence and abundance: a review of focal patch studies. Landscape Ecology 26 7 18. Tillman, E. A., A. van Doorn, and M. L. Avery. (2000) Bird damage to tropical fruit in south Florida. Proceedings of the Wildlife Damage Management Conference 9:4759.

PAGE 154

154 Tim m, R.M. (1991) Commentary on vertebrate pest ecology. Ecology and Management of FoodIndustry Pests. (ed. J.R. Gorham), pp. 323324. Association of Official Analytical Chemists, Arlington, VA. Tisdell, C. ( 1982) Wild Pigs: Environmental Pest or Economic Resource?. Pergamon Press. Sydney, Australia. 445 pp. Tourenq, C., Aulagnier S., Durieux L., Lek, S., Meslard, F. Johnson, A. & Martin, J.L. (20 01) Identifying rice fields at risk from damage by the greater flamingo. Journal of Applied Ecology 38 170179. Tracey, J.P., Bomford, M., Hart, Q., Saunders, G. & Sinclair, R. (2007) Managing bird damage to fruit and other horticultural crops. Bureau of Rural Sciences, Canberra. Australia. 268 pp. Turner H. and D. Firth. 2011. Bradley Terry models in R: The Bradl eyTerry2 package (R package version 0.95). Available from : http://CRAN.Rproject.org/package=BradleyTerry (Last day accessed: July 22, 2011) Volpe, N. & Arambur R.M. (2011) Preferencias de nidificacin de la Cotorra Argentina ( Myiopsitta monachus ) en un rea urbana de Argentina. Ornitologa Neotropical 22, 111119. Walsberg, G. E. (1981) Nest site selection and the radiative environment of the arbling Vireo. Condor 83, 86 88. Walters, C.J (1997) Challenges in adaptive management of riparian and coastal ecosystems. Conservation Ecology [online], 1 1. Available from : http://www.consecol.org/vol1/iss2/art1/ (Last day accessed: July 22, 2011) West, B.C. & Parkhurst, J.A. (2002) Interactions between Deer Damage, Deer Density, and Stakeholder Attitudes in Virginia. Wildlife Society Bulletin 30 139 147. White, T.H.,Brown G.G. & Collazo J.A. (2006) Artificial cavities and nest site sel ection by Puerto Rican parrots: a multiscale assessment. Avian Conservation and Ecology cologie et conservation des oiseaux 1 5. Available from : http://www.ace eco.org/vol1/iss3/art5/ (Last day ac cessed: July 22, 2011) Wiens, J.A. (1989) Spatial Scaling in Ecology. Functional Ecology 3 385 397. Wiens, J.A., Rotenberry, J.T., & Van Horne, B.A. 1987. Habitat occupancy patterns of North American shrubsteppe birds: the effects of spatial scale. Oiko s 48, 132 147. Zaccagnini, M.E. (1998) Evaluacin del dao por aves en cultivos. Manual de Capacitacin sobre Manejo Integrado de Aves Perjudiciales a la Agricultura. (eds E. Rodriguez & M.E. Zaccagnini), pp. 85116. United Nations, Food and Agriculture O rganization (FAO). Binacional Project ArgentinaUruguay. TCP/RLA/8965. 171 pp.

PAGE 155

155 Zaccagnini, M.E. (2006) Impacto del carbofurn para las aves y otra fauna silvestre: anlisis de componentes a tener en cuenta para el manejo integrado del dao de cotorras. Inf orme indito. 11 pp. Zaccagnini, M.E. & Bucher, E.H. (1983) Relevamiento de Problemas Ocasionados por Aves a la Agricultura de la Provincia de Entre Rios INTA, EEA Parana. Unpublished report. Paran, Argentina. Zaccagnini M.E. & Cassani, G.E. (1985) Esti macin de las prdidas ocasionadas por aves granvoras en cultivos de Girasol. Informe indito. 10 pp. Zaccagnini, M.E. & Tate, G. (1991) Evaluacin del Impacto de las aves granvoras silvestres a cultivos agrcolas en Entre Ros: Mdulo Girasol. Informe I ndito. INTA, EEA Paran. Pp. 17. Zaccagnini, M.E. & Tate, G. (1992) Evaluacin del impacto de las aves granvoras silvestres a cultivos agrcolas en Entre Ros: mdulo girasol Convenio INTA Provincia de Entre Ros. Informe indito. 22 p. Zaccagnini, M.E ., Conde, A.A. & Bucher, E.H. (1983) Desarrollo de mtodos de muestreo para la evaluacin de las prdidas por aves granvoras en agricultura. Informe indito. Plan 10:2924. INTA EEA Paran. 9 pp. Zaccagnini, M.E., Conde, A.A. & Dabin, E.L. (1985) Comparaci n de dos mtodos de estimacin y determinacin de muestra ptima para la evaluacin del dao por aves en girasol. XI Conferencia Internacional de Girasol pp. 521 528. Mar del Plata. Argentina. Zaccagnini, M.E. & Dabin E.L. ( 1985 ) Aves granvoras en el c ultivo de girasol: revisin sobre daos y control en distintos pases. Publicacin Tcnica No 11. Produccin Agrcola. INTA, EEA Paran. Entre Ros. Zinn, H.C. & Andelt, W.F. (1999) Attitudes of Fort Collins, Colorado, residents toward prairie dogs. Wildlife Society Bulletin, 27, 1098 1106. Zinn, H. C., Manfredo, M. J., Vaske, J. J. & Wittmann, K. (1998) Using normative beliefs to determine acceptability of wildlife management actions. Society and Natural Resources 11, 649 662.

PAGE 156

156 BIOGRAPHICAL SKETCH Sonia Beatr iz Canavelli was born on 1968 in Paran ( Entre Ros, Argentina) She graduated with a bachelors degree in biology at the Universidad Nacional de Crdoba in 1994. Before finishing, she started working in the Wildlife Department at the National Institu te of Agricultural Technology (INTA), Paran Experimental Station, in projects dealing with bird pest management in Entre Ros and Santa Fe provinces (Argentina). In January 1996, she became involved in field projects about Swainsons hawk ecology in Argentina due to the massive mortalities that occurred from pesticides at that time Between 1998 and 2000, she conducted a m asters program at the University of Florida (UF). Upon her return to Argentina, she helped to design and implement a regional bird monitoring program to monitor the status of bird populations on the pampas region. However, because bird mortalities were occurring because of the illegal use of pesticides to kill bird s that damaged crops, and farmers claims about bird damage to crop s conti nued, Sonia returned to research on bird pest damage. Between 2003 and 2005, Sonia completed course and test requirements for a doctor ate program at UF and i n 2006, back again in Argentina, Sonia started a research program on humanwildlife conflicts, incl uding monk parakeet and eared dove damage to crops In addition to research activities, Sonia teaches courses related to wildlife management and conservation at local universities. Sonia has been married to Carlos Cappellacci since 1998, and both live with their lovely daughter s, Luciana (10) and Eugenia (5), in Paran (Entre Ros, Argentina).