UC Berkeley UC Berkeley Electronic Theses and Dissertations
Title The Landscape Ecology of Pest Control Services: Cabbage Aphid-Syrphid Trophic Dynamics on California's Central Coast
Permalink https://escholarship.org/uc/item/31x6s08m
Author Chaplin-Kramer, Rebecca Ellen
Publication Date 2010
Peer reviewed|Thesis/dissertation
eScholarship.org Powered by the California Digital Library University of California
The Landscape Ecology of Pest Control Services: Cabbage Aphid-Syrphid Trophic Dynamics on California’s Central Coast
by
Rebecca Ellen Chaplin-Kramer
A dissertation submitted in partial satisfaction of the
requirements for the degree of
Doctor of Philosophy
in
Environmental Science, Policy and Management
in the
Graduate Division
of the
University of California, Berkeley
Committee in charge:
Professor Claire Kremen, Chair Professor Nicholas J. Mills Professor Perry de Valpine Professor Mary E. Power
Spring 2010
The Landscape Ecology of Pest Control Services: Cabbage Aphid-Syrphid Trophic Dynamics on California’s Central Coast
© 2010
by Rebecca Ellen Chaplin-Kramer Abstract
The Landscape Ecology of Pest Control Services: Cabbage Aphid-Syrphid Trophic Dynamics on California’s Central Coast
by
Rebecca Ellen Chaplin-Kramer
Doctor of Philosophy in Environmental Science, Policy and Management
University of California, Berkeley
Professor Claire Kremen, Chair
Agricultural pests can be reduced or controlled by naturally-occurring predators and parasitoids, but effective management of this ecosystem service requires an understanding of the factors contributing to its delivery. Pest control services are thought to be enhanced by natural habitat in agricultural landscapes, based on a large body of research demonstrating positive relationships between predator or parasitoid abundance on farms and the amount of nearby natural habitat. However, there is little or no evidence for a concomitant increase in pest control—if “control” is defined as the maintenance of pest populations below a certain level. The difficulty in measuring pest control by pest densities is that it is unknown how much more abundant pests would be in the absence of their enemies. This dissertation investigates different aspects of pest control to assess whether and under what conditions natural habitat can provide this ecosystem service to farms, using the study system of cabbage aphids (Brevicoryne brassicae) and their syrphid predators (Diptera: Syrphidae) in broccoli (Brassica oleracea) on California’s Central Coast.
A major factor that may affect the potential for natural habitat to provide pest control services is the presence of alternate host plants for pests within the weed community surrounding the farm. Physiological experiments show that one weedy relative of broccoli, black mustard (Brassica nigra), could serve as more than merely an alternate host. Due to its toxic chemistry, B. nigra allows cabbage aphids to develop more quickly than when on broccoli, and to build a mustard bomb that can compromise or kill their syrphid predators. Weedy B. nigra may provide a refuge from an important predator and thus could be serving as a source of aphid pests to crops. The potential for habitat around the farm to benefit pests as well as natural enemies is an important consideration in understanding the impact of habitat complexity on pest control services.
Underlying spatial variation in pest distributions (whether due to alternate host plants or a number of other factors such as microclimate and dispersal) makes it difficult to detect the contribution of natural enemies to overall pest control. A cage experiment holding initial aphid densities constant across a landscape gradient measures the effect of habitat
1 complexity on pest reduction by natural enemies, and reveals the importance of habitat at the landscape level as well as at the local on-farm level late in the growing season. Reduction of cabbage aphid densities is up to four times higher at more complex sites. The degree to which habitat at one scale could compensate for a lack of it at the other is dependent on aphid colonization and population growth early in the season, determined by abiotic factors such as temperature and dispersal patterns.
In addition to considering such variation across spatial scales, this dissertation tackles variation in pest populations and the delivery of pest control services across temporal scales. A three-year survey involving weekly measurements of aphid and syrphid densities assesses the effect of natural habitat on different indicators of pest control services. Syrphid larval abundance increases strongly with the proportion of natural habitat surrounding the farm, and cabbage aphid population growth is reduced on farms with higher syrphid densities, resulting in a weak but significant decline of aphid densities with natural habitat. Pest control services delivered by natural habitat may be masked by inter-annual variation in environmental factors and by competing direct and indirect effects (i.e., landscape effects on the pests themselves versus on the natural enemies’ control of pests). Therefore, longer term datasets are necessary in order to detect the true magnitude of pest control services provided by natural habitat.
The results of this dissertation indicate that pests are indeed constrained to some extent by ecosystem services, but the degree to which this constraint can contribute to overall pest control is dependent on pest colonization and population growth early in the season. The role of natural habitat in promoting natural enemy communities is an important consideration for pest management, but must be employed with other methods of pest management for truly effective and reliable control.
2 Table of Contents
List of Tables ...... iii List of Figures...... iv Acknowledgements...... v
Chapter 1: Introduction...... 1 Can habitat complexity provide pest control services to farms? ...... 2 What aspects of habitat complexity compromise the delivery of pest control services to farms? ...... 2 What is the role of different scales of habitat complexity in providing pest control services on farms?...... 3 How effective are pest control services at suppressing pest densities on farms? ...... 3
Chapter 2: Invasive Mustard Promotes Aphid Growth While Eliminating Syrphid Control...... 4 Introduction...... 5 Materials and Methods...... 6 Chemical assays...... 6 Aphid development and fecundity ...... 7 Predation and predator development ...... 7 Analysis...... 8 Results...... 9 Chemical assays...... 9 Aphid development and fecundity ...... 9 Predation and syrphid development...... 9 Discussion ...... 14 Conclusion ...... 16
Chapter 3: Pest Control Experiments Show Benefits of Complexity at Landscape and Local Scales...... 18 Introduction...... 19 Materials and Methods...... 21 Study system...... 21 Cage Experiment...... 24 Analysis...... 26 Results...... 27 Discussion ...... 34 Conclusion ...... 36
i Chapter 4: Subtle Effects of Pest Control Services Along a Landscape Gradient and Across Temporal Scales ...... 38 Introduction...... 39 Materials and Methods...... 41 Study system...... 41 GIS land-use mapping...... 42 Insect collection...... 42 Analysis...... 46 Results...... 47 Discussion ...... 50 Conclusion ...... 56
Chapter 5: Conclusion...... 58 Main findings ...... 58 Future directions ...... 60 Implications of the research ...... 60
References ...... 61
ii
List of Tables
2.1a Glucosinolate content of host plants, mustard (Brassica nigra) and broccoli (B. oleracea)...... 10 2.1b Glucosinolate content of cabbage aphids reared on mustard vs. broccoli...... 10 2.2 Lifetime aphid consumption, average daily aphid consumption, and maximum daily aphid consumption by three syrphid species...... 13 3.1 Study sites, characterized by local complexity and landscape matrix, with % natural habitat within 2.5 km of the farm...... 23 3.2 ANOVA for effects of landscape matrix and local complexity on proportional reduction in aphid densities...... 29 3.3 ANOVA for effects of landscape matrix, local complexity, and season on pest pressure for 2009...... 30 3.4 The linear model of field aphid densities explained by PRD and R...... 33 4.1 Study sites, according to years included in study and % natural habitat in 2.5 km.....44 4.2 The relationship between syrphid abundance and % natural habitat at different scales ...... 48 4.3 Models explaining variation in MV∆...... 53 4.4 Models explaining variation in final aphid densities...... 54
iii
List of Figures
2.1 Impacts of host plant (broccoli vs. mustard) on cabbage aphid physiology...... 11 2.2 Impacts of cabbage aphid host plant (broccoli vs. mustard) on syrphid larvae...... 12 3.1 Map of study sites, in the Salinas Valley and surrounding areas...... 22 3.2 The effect of landscape matrix and local complexity on the proportional reduction in aphid densities (PRD)...... 28 3.3 Pest pressure for 2009, broken down by colonization and reproduction...... 30 3.4 The effect of season (early or late) and maximum temperature (°C) on aphid net reproductive rate (R) in 2009...... 31 3.5 The effect of aphid net reproductive rate (R) and the presence of mustard on aphid colonization rate (C) in 2009...... 32 4.1 Map of study sites, in the Salinas Valley and surrounding areas...... 43 4.2 Log aphid densities at all field sites over the summer season (May – September).....45 4.3 Effect of proportion of natural habitat within 2.5 km on mean syrphid abundance....49 4.4 Effect of previous week’s syrphid densities on change in aphid densities...... 51 4.5 Effect of year and previous week’s syrphid densities on minimum value of change in aphid densities...... 52
iv
Acknowledgements
This work was funded through the generous support of The Environ Foundation, The Land Institute, the USDA’s Sustainable Agriculture Research & Education Program, The Organic Farming Research Foundation, The Van Den Bosch Foundation, and the National Science Foundation’s Graduate Research Fellowship Program.
I would first like to thank the growers who gave me both access to their farms for my research and an enthusiastic reception whenever I shared with them what I had learned there. Ramy Colfer at Mission Organics, Mark Marino at Earthbound Farms, Greg Rawlings at Blue Moon Organics, Ron Yokota, Mike Thorp, and Scott Rossi at Tanimura & Antle, Alan Miramura at Lakeside Organics, Tim Campion at Swanton Berry Farm, Liz Milazzo and Jim Leap at the UCSC Center for Agroecology and Sustainable Food Systems, Sharon Benzen and Jim McCreight at the USDA, Bill Sullivan at Crown Packing, John Savage at Pure Pacific, all the staff at ALBA Farms, and Phil Foster at Pinnacle Organics—I am grateful for their patience with me and their willingness to take a chance on unusual research without an immediate promise of pay-off. They might have wondered how deliberately putting pests out on their farms was going to help their pest problems, but they let me do it anyway. Grower’s Transplanting sold me broccoli one flat at a time, which was almost certainly not an economy of scale for them, but I sure appreciated it. Over half those connections I owe to my friend and colleague Sara Bothwell. Enormous thanks go to her for introducing me to the system and showing me around the agricultural side of the Central Coast.
So many students came out with me to those farms, to get stuck in the mud or bake in the sun or just drive around in a car reeking of sweaty broccoli. They never complained and many hands made light work, both in the field and back in the lab. The sweat and tears of Kanthinh Manivong, Naomi Schowalter, Andrea Chiem, Elizabeth Morrill, Grace Chuang, Addie Abrams, Giancarlo Leonio, and Jennifer Bartlau run throughout this dissertation.
The greenhouse staff at the Oxford Tract provided more invaluable assistance. Barb Rotz and John Franklin kept my plants and aphid colonies alive, and helped me weather many an unwanted outbreak from the wrong species of aphid. John Andrews gave me a crash course in just about everything to do with aphids; I was so lucky to have him as a constant resource for trouble-shooting and advice. I thank Dan Kliebenstein for his help with the glucosinolate analyses, and for taking me on as a collaborator without ever even having met me. His quick replies to my unending questions on a subject I knew so little about were what made the second chapter possible. And the mustard bomb phenomenon would never have gotten on my radar if not for a well-timed article sent to me by Jeff Lozier with the subject line “did you know your aphids are eco-terrorists?”
v
I also relied heavily on the support staff of the Geospatial Innovation Facility at Berkeley, especially Karin Tuxen-Bettman, who gave me a better education in GIS and remote- sensing than any class or manual could do, who always made time to answer my questions or help me find my way around a bug in a program, and who still responds to my emails even though she doesn’t even work there anymore. Karin makes landscape ecology exciting and accessible and I hope Google Earth knows how lucky they are to have snatched her up.
My dissertation committee was a fantastic mix of people. Nick Mills gave me a sounding board for all my ideas about pest-predator dynamics, and always had a paper to recommend that would put it in perspective for me. His pragmatism helped me stay on track, a nice counterpoint to the imaginative and ingenious Mary Power. Her encouragement was as effusive as her myriad ideas for all the directions I could take things. Mary makes me feel like there really is no end to any research question, just lots and lots of different beginnings. Perry de Valpine was my analysis guru all along the way, from the big picture level of my experimental design to the nitty gritty details of my R code. He is so mild-mannered about his mathematical brilliance; I am so grateful to have that statistical superman on my side. And finally Claire Kremen, my major professor, my hero. She brings insightful questions, incisive criticism, and inspiration above all else. I can’t thank her enough for the scientist she has helped me become, and I look forward to a long future of collaborations.
I thank my labmates in the Kremen lab, as well. Claire has cultivated quite the atmosphere of excellence, and I am humbled to be a member of such a group. Alex Harmon-Threatt, Chris Golden, Cause Hanna, Mez Baker, Brian Gross, Tendro Ramaharitra, Hillary Sardinas… you guys are rockstars. Thank you especially to our post-docs Lora Morandin and Shalene Jha, for their editing help and good career advice. I don’t feel smart enough to be so close to being in your shoes.
I am very grateful for the friendships I have formed in our department, from those first few weeks in our cohort, to what I know I will take with me the rest of my life. EJ Blitzer, Kayje Booker, Emily Rubidge, Carrie Cizauskas. Thank you for your humor, your confidence, your passions, and your example. These are some pretty amazing women.
Finally, I am so thankful for the love and support of my family. My parents were my very first role-models of scientists, and I know I owe this path I’ve been on for most of my life to them. Thank you for the camping trips, the hikes, even the boring bird-watching. Did you ever imagine the apple would fall so close to the tree? I am so lucky to have a whole family willing to accompany me into the field—my parents, my grandma, a whole summer’s worth of employment out of my husband… thank you for coming along for the ride in more ways than one. I am indebted to my parents, my parents-in-law, and especially my grandma for unbelievable amounts of free childcare while I madly finished up this research. The secret to having a full-time career and being a full-time mom is having lots of full-time family. And to my dear husband, Danny, who has put up with all
vi my I-just-need-a-few-more-minutes that inevitably always turned into many, many hours: thank you for your understanding, your support, your love, your picking up the slack, your much-needed diversions, your company on long trips, your handy-ness with a hammar and saw for field construction projects, your eye for policy implications in my grant proposals and reports, and most of all your unflinching and unfailing belief in me.
And to Jack. Thank you for letting me drag you along on all my field trips. Thank you for your good nature, and thank you for giving me the best inspiration of all: you are the reason I want to make this world a better place.
vii Chapter 1: Introduction
The intensification and expansion of agriculture in the latter part of the twentieth century has amplified competition between humans and arthropod herbivores for food produced from crops. Despite soaring production costs and serious threats to human and ecosystem health, synthetic pesticides have been used as the main line of defense against arthropod pests since the 1940s. Humans could reduce reliance on pesticides, while providing a natural and self-sustaining means of pest control, by enhancing natural populations of insects that prey upon agricultural pests.
Pest control by chemical means is neither economically nor ecologically sustainable. The U.S. spent an estimated $11 billion on pesticides in 2008, applying more than 480 million pounds of these chemicals to its agricultural acreage (Fernandez-Cornejo et al. 2010). Even with the use of pesticides, it has been estimated that 37% of crop yields are lost to pests (Pimentel et al. 1992). Pesticides pose myriad threats to humans and wildlife; some pesticides are carcinogenic and others disrupt endocrine and nervous system function, leading to complications in reproduction and/or serious illness (IPCS 2002, PANNA 2005). The irony of pesticides is that they can exacerbate the pest problems they were designed to solve. Pesticides can kill non-target enemies of pests, thus leading to secondary pest infestations when pest populations rebound faster than natural enemy populations or when formerly unimportant herbivores are released from control (Kenmore 1980, Settle et al. 1996). The pests themselves can escape pesticide control by rapidly evolving resistance to pesticides, making these chemicals less effective with repeated use (Naylor & Ehrlich 1997). A more sustainable solution to the problem of agricultural pests lies in a natural agent that can evolve along with the pest, that can respond to pest population dynamics, and that produces no contaminating pollutants.
Predators and parasitoid insects can provide an ecosystem service to agriculture by suppressing or reducing pest populations on farms. This ecosystem service has an estimated value of $13 billion per year in the U.S. (Losey & Vaughan 2006), but could be greatly enhanced if we better understood the mechanisms contributing to its delivery. Most natural enemies require off-farm habitat for nectar, alternate prey, and/or overwintering sites, making the landscape composition around the farm an important factor in the provision of pest control services (Tscharntke et al. 2007). Landscape complexity as measured by the proportion of natural or non-crop habitat around farms has been widely shown to increase natural enemy abundance, diversity and function (predation or parasitism). However, the relationship between landscape complexity and pest densities or crop damage remains inconclusive (Bianchi et al. 2006).
1 The dissertation research presented here investigates the conditions under which top- down control by natural enemies can provide effective pest control on farms, and identifies factors that may mask the role that natural enemies play in determining pest population dynamics. The study system of broccoli (Brassica oleracea), cabbage aphids (Brevicoryne brassicae), and their syrphid predators (Diptera: Syrphidae) presents an opportunity to explore research questions in a fairly simple food web. Although there are other pests (notably the lepidopterans Pieris rapae and Trichoplusia ni) and a broader suite of natural enemies in this system, aphids and syrphids are the predominant players. The overarching question explored in this dissertation is:
Can habitat complexity provide pest control services to farms?
This question, while the most basic, is also the most difficult to answer, and depends in part on how pest control is defined. Traditionally, “pest control” is defined by economic thresholds and plant damage. “Pest control services” could be considered more broadly, as the contribution made by natural enemies to buffer the crop against pest pressure. From this overarching focus, secondary questions arise and must be addressed in order to understand the role of natural habitat in pest-predator dynamics. This first question both precedes and concludes the remaining questions, growing more nuanced as the intricacies of pest-predator spatial and temporal dynamics are uncovered.
Before considering the ecosystem services that habitat complexity can provide to farms, it is important to also recognize the potential for ecosystem disservices. One concern is that natural areas provide habitat for pests as well as enemies. This leads to the question:
What aspects of habitat complexity compromise the delivery of pest control services to farms?
Chapter Two investigates this question by testing the impact of alternate host-plants on pest and predator demography through a series of physiological laboratory experiments. A major factor that may affect the potential for natural habitat to provide pest control services is the presence of alternate host plants for pests within the surrounding weed community. These weeds could make natural habitat as much a source of pests as of enemies. In this system, a weedy relative of broccoli, black mustard (Brassica nigra), may in fact be more than just an alternate host; it may provide enemy-free space near farms by allowing aphids to build up their populations on a more toxic food source before dispersing to the crop plants. The hypotheses tested are: 1) aphids feeding on mustard can better avoid predation than aphids feeding on broccoli, and 2) aphids feeding on mustard pay a physiological cost for that protection. The goal of this study is to weigh the costs and benefits of mustard to aphids in order to assess its potential to serve as a source of crop pests.
The presence of source habitat for pests off the farm is one factor that can influence the provision of pest control services by natural habitat, but the farm habitat itself can also play a role. To the extent that the farm itself provides resources for natural enemies,
2 these local resources may enhance or compensate for natural habitat at a landscape scale. This lead to the question:
What is the role of different scales of habitat complexity in providing pest control services on farms?
Chapter Three addresses this question through a controlled field experiment to measure the reduction of pest densities by natural enemies on farms differing in both local (on- farm) and landscape complexity. This experiment determines the effect of complexity on pest control services across scales, teasing apart the relative contributions of landscape and local complexity. Furthermore, the experimental design offers insight into the role of temperature and the presence of source habitat for pests in mitigating these effects. The hypotheses tested here are: 1) habitat complexity at both landscape and local scales provides better pest control services than complexity at only one scale, and 2) complexity at one scale can compensate for the lack of it at another. The goal of this chapter is to determine the importance of different scales of habitat complexity to pest control services and conditions under which those services are maximized.
Since pest control services are the contribution made by natural enemies to buffer the crop against pest pressure, the issue remains whether “pest control service” actually controls pests. The final question of this research is:
How effective are pest control services at suppressing pest densities on farms?
Chapter Four tackles this question by connecting landscape-mediated increases in natural enemy abundance to the resulting impacts on pest populations. Syrphids may be reducing aphid populations from what they would be in absence of top-down control, without fully suppressing their growth or maintaining them below a desired level. Weekly insect surveys over three years and across a landscape gradient test the hypothesis that natural habitat increases syrphid abundances on farms, suppresses aphid population growth, and results in lower aphid infestations of broccoli. The goal is to determine to what degree natural habitat can deliver enemies to provide truly effective and reliable control. This brings us full circle to the initial question, whether habitat complexity provides pest control services to farms, with implications for the conservation of natural habitat as a pest management strategy.
3
Chapter 2: Invasive Mustard Promotes Aphid Growth While Eliminating Syrphid Control
ABSTRACT: Specialist herbivores such as the cabbage aphid (Brevicoryne brassicae) co-opt the defense system of brassicaceous plants by sequestering glucosinolates to utilize in their own defense. Cabbage aphids growing on high-glucosinolate plants like Brassica nigra contain more glucosinolates than aphids on cultivated lower-glucosinolate Brassica oleracea. Aphids on B. nigra reproduce more quickly, and compromise or kill their main predators, syrphid larvae (Diptera: Syrphidae). This study presents evidence that weedy B. nigra provides enemy-free space from an important predator and the possibility that this plant could be serving as a source of aphid pests to crops.
4 Introduction
Plants can provide an ecological refuge for herbivores by allowing them to chemically or physically escape their natural enemies, sometimes called an ‘enemy-free space’ (Jeffries & Lawton 1984). Although the best examples of enemy-free space have been documented in natural systems and through experimental host-plant shifts (Denno et al. 1990, Gratton & Welter 1999, Murphy 2004), the concept has useful application in the context of biological control for agricultural systems. Identifying plants that could provide enemy-free space for herbivorous insects in different cropping systems would help guide measures to reduce or eliminate an important source of pests to crop fields. Many wild plants contain higher levels of defense compounds than their domesticated congeners, and specialized herbivores can often utilize those toxins to compile their own chemical arsenal to escape enemies. If this chemical refuge allows pests to more effectively escape their enemies than when feeding on crops, the occurrence of such plants around farmland could be supporting populations of pests without a substantial enemy community. Such a refuge could thus give pests a head start on colonizing the crops before their enemies arrive. The invasive mustard Brassica nigra is commonly found in field margins or edges around domesticated Brassica oleracea fields (cole crops, such as broccoli, cauliflower, kale, cabbage). The co-occurrence of these two Brassica species provides an excellent opportunity to test whether an invasive weed can serve as enemy-free space for crop pests.
Plants in the family Brassicaceae are a good study system for chemically mediated enemy-free space, due to their sophisticated two-part defense system involving a glucosinolate compound and myrosinase protein complex that has been described as a “mustard-oil bomb” (Ratzka et al. 2002). The cabbage aphid is among a small group of insects that have found a strategy for exploiting Brassicaceae, a resource toxic to most herbivores. The handful of generalist herbivores adapted to the glucosinolates in these plants tend to rely on detoxification, receiving little or no predator defense benefit (Hopkins et al. 2009). In contrast, specialist herbivores co-opt toxicity from glucosinolates through excretion or sequestration, escaping many of their enemies as a result (Hopkins et al. 2009, Muller et al. 2001, Aliabadi et al. 2002, Rayor et al. 2007, Harvey et al. 2007). What makes the cabbage aphid relatively unusual among these specialists is the pairing of a myrosinase complex with the glucosinolate sequestration, such that they not only deter but actually harm their predators (Hopkins et al. 2009). The aphids avoid the toxic products resulting from the degradation of these compounds by rapidly partitioning the glucosinolate into a separate sub-cellular compartment and producing their own myrosinase enzyme, essentially arming themselves with their own “mustard bomb” (Bridges et al. 2002). When the aphid body is damaged the two compounds are mixed and the hydrolysis of glucosinolate by the myrosinase produces volatile isothiocyanates (Francis et al. 2004), which are directly toxic to predators (Francis et al. 2001).
The degree to which a high-glucosinolate plant like B. nigra serves as an enemy free space, and thus a source of aphids to nearby cropland, depends on the tradeoff between enemy-free space and aphid development. Cabbage aphids are completely dependent on 5 their food source for the acquisition of the mustard bomb. They cannot synthesize their own glucosinolates (Kazana et al. 2007), and therefore can only create the bomb to kill or deter their natural enemies if they consume a glucosinolate diet (Olmez-Bayhan et al. 2007, Francis et al. 2000, Pratt et al. 2008, Vanhaelen et al. 2002). Thus, differences in the concentration of glucosinolates within the aphids’ diet may influence their ability to deter predation. However, any benefit of predator refuge derived from glucosinolates may be offset by slower aphid development or reproduction, due to potential energetic costs of sequestering these toxic compounds. While not framing their study specifically in terms of glucosinolate content, Ulusoy and Olmez-Bayhan (2006) found that aphids grown on another wild mustard, Sinapis arvensis, had lower reproductive rates than those reared on B. oleracea cultivars. Aphids on S. arvensis also had shorter generation lengths, however, resulting in no net difference in intrinsic rate of increase for aphids on wild and domesticated mustards in the absence of predation.
If wild Brassica species provide enemy-free space without compromising a pest’s own population growth, it is possible that these plants serve as sources of pests to nearby crops. Evidence for this possibility is lacking in the literature, since herbivore development studies have not been conducted on the same plant species as the studies investigating effects on their predators. In order to assess the trade-offs of glucosinolate sequestration, if any exist, this study measured both herbivore development and predation in the same system. Specifically, we quantified the glucosinolate content of two Brassica species (B. oleracea and B. nigra), measured the glucosinolate content in turn of aphids reared upon the two plant species, assessed aphid development and fecundity while growing on these two food sources, and evaluated syrphid development and predation of aphids raised on each host plant. We tested the hypotheses that: (1) high glucosinolate content puts a physiological burden on the herbivores utilizing these compounds for their own defense, and (2) high glucosinolate content confers an advantage to the herbivore in the form of reduced predation. This laboratory research highlights the impact of phytochemicals on pests and their predators, with implications for the role an invasive weed may play in inhibiting effective biological control in agricultural systems.
Materials and Methods
Chemical assays Aphids (Brevicoryne brassicae) were reared for a period of three months in dense (> 100 per leaf) colonies on broccoli (Brassica oleracea) and black mustard (Brassica nigra) at 18 - 24ºC. Apterous aphids (those lacking wings) were selected from six B. oleracea and six B. nigra plants and tested for glucosinolate content. Samples consisting of ten aphids in their penultimate instar were taken from each individual plant, and preserved in 500 µl of 90% methanol for glucosinolate analysis. Fresh leaf matter totaling150-250mg was also collected from each individual plant in each treatment for analysis.
Glucosinolates were measured via High-Performance Liquid Chromatography (HPLC) with DAD detection using an Agilent 1100 system as described in Kliebenstein et al. 6 (2001). Individual glucosinolates detected were identified and quantified in relation to previously purified standards (Reichelt et al. 2002). After quantification with regards to each individual glucosinolate’s purified standard, all individual aliphatic glucosinolate values within a sample were summed to develop a measure for total aliphatic glucosinolate within a sample and standardized to nmol per unit fresh weight.
Aphid development and fecundity Small cages were clipped directly onto leaf surfaces of potted B. oleracea or B. nigra plants in the greenhouse, at 18 - 24ºC, and 16-hour day-length. Each clip-cage contained a group of several adult aphids (denoted hereafter as the zero generation, or G0). The clip-cages allowed individual aphids or offspring cohorts to be tracked on living plants rather than detached leaves; this is important for simulating true field conditions, since plants will often increase the amount of glucosinolates in the leaf when they are being attacked (Cui et al. 2002, Kim & Jander 2007).
Daily observations were made of the clip-cages, so that the first day that nymphs were observed on the leaf, all the G0 adults were removed and the cohort of nymphs (G1) was tracked until first reproduction. Once the next generation of nymphs (G2) was produced, each of the G1 aphids (now adults) was moved to its own cage on a separate leaf and tracked individually. This experimental design reduced aphid escape from the cages, since nymphs born together tend to remain where they are on the leaf, while individual nymphs will often search the plant for other aphids (personal observation). Once the G1 adults were in their own individual cages, they continued to be checked on a daily basis for nymph production. The initial reproduction event that prompted the movement of G1 aphids to their individual cages could not be attributed to a particular adult, so one day of data is missing for one or more adults, but this bias was conserved across treatments and should not affect comparative results. The number of days until the first G2 nymph was produced was recorded for each G1 individual as a measure of their development time, from birth to reproductive maturity. The number of nymphs produced per G1 aphid each day provided a measure of reproductive rate and total fecundity, as the G1 aphids were tracked until their death. The G2 nymphs were removed each day after counting to remove any artifact of crowding from the effect of the cage.
Forty G1 individuals were tracked for each treatment, but some aphids escaped from the cages during the experiment, since the leaf surfaces were too irregular for the cages to make a perfect seal. Individuals that escaped before reproduction were excluded from the analysis. Individuals that escaped after reproduction were included in the development (time to first reproduction) analysis, but excluded from the analyses on total fecundity. This approach yielded 59 G1 aphids (28 in the B. oleracea treatment and 31 in the B. nigra treatment) for the development analysis, and 32 G1 aphids (17 in the B. oleracea treatment and 15 in the B. nigra treatment) for the fecundity analysis.
Predation and predator development Syrphid larvae are the most common predators of cabbage aphids feeding on B. oleracea (Chapter 4), and are thought to play a vital role in their control for this important crop.
7 Syrphid larvae were reared on aphids from the two different Brassica food sources and observed over the course of their larval development in a growth chamber at 18ºC and a 16-hour daylight photoperiod. Syrphid eggs were collected from the field on B. oleracea plants at the University of California Center for Sustainable Agriculture and Food Systems, in Santa Cruz, CA. B. nigra was also searched for syrphid eggs at the same site, but no eggs were found on these plants. The syrphid species found at this site included Allograpta obliqua, Eupeodes americanus, Eupeodes volucris, Syrphus opinator, Scaeva pyrastri, Sphaerophoria sulphuripes, Toxomerus occidentalis, and Platycheirus stegnus (Chapter 4). As species cannot be identified at the egg stage, it was not possible to ensure an equal number of each species was assigned to each treatment, but the distribution of species was presumably random across the two treatments.
Twenty syrphids were tracked in each treatment from hatching until the larva died or pupated, and the emerging adults were saved for later species identification. The eggs were kept in a moist Petri dish and checked several times per day for hatching. Immediately after hatching, each syrphid larva was placed in a Petri dish on either a B. oleracea leaf or a B. nigra leaf, with aphids from colonies reared on the corresponding plants. Petri dishes with cut leaves were used in this case because clip-cages could not contain the first-instar syrphid larvae. The senescence of the cut leaves in the Petri dishes may have reduced the glucosinolate content of the food source that the aphids were feeding on immediately prior to the introduction of the syrphid larva. However, this period represented a very small fraction of the aphids’ lifetime of feeding, and any reduction in aphid glucosinolate levels should diminish the difference between treatments, providing a conservative estimate of the impact of prey food source on predation.
Only pre-reproductive aphids in their penultimate instar were selected as prey, to standardize feedings and to ensure that no new nymphs were born between feedings. Each day, the number of aphids remaining was recorded, aphids were replenished to ensure a sufficient amount for the syrphid to reach satiation (ranging from 10 to 200 aphids, depending on the age of the syrphid), and the leaf was replaced to maintain freshness. Aphid consumption rate and total lifetime consumption were used as measures of predation by syrphids on the two different aphid sources. The number of days until pupation and mortality provided a comparison of syrphid development in the two treatments.
Analysis All analyses were performed using the statistical program, R (version 2.9.1, http://www.R-project.org). Differences in syrphid mortality between B. oleracea and B. nigra treatments were compared using a χ2 -squared test. All other pairwise comparisons were analyzed using Analysis of Variance.
8 Results
Chemical assays Analysis of B. nigra and B. oleracea plant material by HPLC confirmed previously reported differences in plant glucosinolate profiles (Mithen et al. 1995, Rodman et al. 1981). B. nigra contained several orders of magnitude higher total aliphatic glucosinolate concentration (Table 2.1a, 155 ± 46 nmol /100mg fresh plant matter, mean ± standard error) than B. oleracea (0.68 ± 0.15 nmol/100mg fresh plant matter). Mirroring the endogenous glucosinolate differences between the two food plants, aphids reared on B. nigra sequestered nearly ten times the aliphatic glucosinolates of aphids reared on B. oleracea (Table 2.1b, 17.99 ± 4.72 nmol/10 aphids and 1.89 ± 0.37 nmol/10 aphids, respectively; F = 11.56, df = 1,10, p = 0.007). This suggests that the aphid’s ability to sequester glucosinolates is not saturated when reared upon B. oleracea and that the food source can determine the aphid’s endogenous defense capacity.
Aphid development and fecundity Aphids reared on B. nigra reached reproductive maturity 14% faster on average than those reared on B. oleracea (Figure 2.1a, F = 18. 84, df = 1, 57, p < 0.001). Aphids in the B. nigra treatment also reproduced at a faster rate, producing just over 3 nymphs per day, compared to 2 nymphs a day for aphids on B. oleracea (Figure 2.1b, F = 10.84, df = 1, 57, p = 0.003). However, aphids survived significantly longer on B. oleracea than B. nigra (Figure 2.1c , F = 7.46, df = 1, 28, p = 0.011), resulting in no significant differences in overall nymph production between the two treatments (Figure 2.1d, F = 0.067, df = 1, 57, p = 0. 79).
Predation and syrphid development Syrphid larval mortality in the B. nigra treatment was 95%, more than double that found in the B. oleracea treatment (Figure 2.2a, χ2 = 9.64, df = 1 p = 0.002). Only one larva in the B. nigra treatment pupated, and that individual never emerged from its pupa. As a result of this high mortality, syrphids in the B. nigra treatment spent a mere 10.5 days as larvae (and hence as predators) compared to an average of 17 days in the B. oleracea treatment (Figure 2.2b, F = 5.26, df = 1, 38, p = 0.027). Syrphids in the B. oleracea treatment consumed aphids 65% faster than syrphids in the B. nigra treatment (Figure 2.2c, F = 4.25, df = 1, 38, p = 0.046). Due to the substantial impact of B. nigra on syrphid larval development, the difference between the two treatments in total lifetime aphid consumption by syrphid larvae was even greater with overall predation in the B. oleracea treatment being 200% of that in the B. nigra treatment (Figure 2.2d, F = 4.18, df = 1, 38, p = 0.048). Of the 12 larvae that reached pupation, two did not emerge from their pupae. Of the remainder, 50% were Allograpta obliqua, 30% were Sphaerophoria species, and 20% were Eupeodes americanus (Table 2.2).
9
10
11
12
13
Discussion
Cabbage aphids feeding on B. nigra contain more glucosinolates than aphids feeding on B. oleracea. Aphids reared on B. oleracea still have detectable levels of glucosinolates, but it appears that the concentrations of these compounds are not high enough for aphids to effectively deter their enemies. Other studies have also demonstrated that a certain threshold of glucosinolate content must be reached before aphids can sequester enough to deter or damage their enemies. The coccinellid Adalia bipunctata was able to complete its larval development on aphids reared on Brassica napus, which is lower in glucosinolate content, but not on aphids reared on the more concentrated glucosinolates of Sinapis alba (Francis et al. 2000). Other natural enemies show slower growth rates or reduced reproduction rather than direct mortality, but similarly seem less impacted by glucosinolates below a certain threshold (Vanhaelen et al. 2002, Olmez-Bayhan et al. 2007). While the number and variety of specific glucosinolate compounds prevents a direct comparison between this study and previous work detecting these threshold effects, the general trend is congruent with the findings of this study: low-glucosinolate plants can pass some of these compounds onto herbivores without providing the same protection from predators offered by high-glucosinolate plants.
Surprisingly, we did not measure any physiological cost to aphids using B. nigra chemicals as ammunition against their enemies. The sequestration of glucosinolates may not impart a significant energy cost upon the aphid as this system only requires a transporter to move the glucosinolate out of the digestive stream. Recent work has suggested that cabbage aphid may avoid the myrosinase cells via a behavioral change in stylet penetration, reducing or eliminating energetic cost (Kim & Jander 2007). The lack of an energetic cost of glucosinolate sequestration does not explain the faster growth rate we found for aphids on the high-glucosinolate B. nigra, however. It is possible that development is accelerated because overall weight gain is reduced, as has been shown to be the case for other herbivores and increasing aliphatic glucosinolate content (Beekwilder et al. 2008). If this were the case, overall fecundity should be lower, because smaller aphids have been shown to have fewer and smaller offspring (Dixon 1997). However, aphid fecundity was not significantly different between the two treatments in this experiment; a longer lifespan and thus more days of reproduction on B. oleracea was countered by a faster reproductive rate on B. nigra.
The question of how glucosinolates impact cabbage aphid physiology is complex, because Brassica species or cultivars have different glucosinolate profiles, which, in turn, have different physiological consequences for herbivores. Cole (1997) found an indolic glucosinolate suppressed aphid intrinsic rate of increase, while two aliphatic glucosinolates enhanced it. Cabbage aphids in Cole’s study had faster development and lower fecundity on B. nigra than on most of the B. oleracea cultivars, but there was one cultivar of B. oleracea (a cauliflower) on which aphid fecundity was not markedly different than on B. nigra. Broccoli was not one of the cultivars tested by Cole, but our glucosinolate profile showed that it contained markedly less sinigrin, one of the aliphatic
14 glucosinolates found to be positively correlated with intrinsic rate of increase of cabbage aphid. While the results presented here may not be generalizable to all cultivars of Brassica oleracea, the fact that aphids reared on B. nigra in our study reached reproductive maturity faster without experiencing a decline in total lifetime fecundity suggests that overall aphid production at the population level is enhanced on B. nigra compared to broccoli. Faster generation times with equivalent reproductive rates translate into faster population growth (Stearns 1992). Rather than a physiological cost of living on B. nigra, there appears to be a slight benefit, even beyond that provided by potential enemy escape.
Our results show that syrphids are far less effective predators of cabbage aphids on B. nigra than on B. oleracea. They consume aphids more slowly and far fewer aphids total on B. nigra. Moreover, the dramatic larval mortality observed in these trials suggests that there could be heavy selection pressure against ovipositing on B. nigra. Rather than experiencing this kind of mortality in the field, it is possible that adult syrphid flies simply avoid B. nigra altogether. Supporting this speculation is our observation that while syrphid fly adults feed at B. nigra flowers and were often seen flying around the plant, we never found syrphid eggs on wild B. nigra.
Field studies are necessary to ascertain to what degree B. nigra truly provides enemy-free space, as suggested by this physiological work. Future studies should determine the extent to which other enemies affect aphid distributions on B. nigra in the field and test the potential for patches of B. nigra to export aphids to neighboring brassicaceous crops. Comparing abundance and composition of natural enemies of aphids on cultivated and wild Brassicas will help to identify whether different players are more prominent in these two systems. Parasitism rate is clearly an important factor to consider, since specialist parasitoids may not be as vulnerable to the aphids’ mustard bomb as are the generalist predators. Parasitism of aphids is very low in the domesticated Brassicas in this system (<10%, Chaplin-Kramer unpublished data), and is unlikely to provide effective control at such a rate (Hawkins & Cornell 1994). However, the role parasitoids play in wild Brassicaceae could be very different.
It is possible and quite likely that other natural enemies are preying on cabbage aphids on B. nigra in the field. Parasitized aphids have been observed on the stems of these plants, suggesting that parasitoids still play an important role in influencing aphid populations on B. nigra. Their evolutionary relationship with these aphids as specialists may better equip them to manage the higher glucosinolate concentrations than generalist predators such as syrphid larvae. In fact, parasitoids attacking generalist herbivores are more affected by toxins in the herbivore’s diet than parasitoids attacking herbivores specialized on Brassicaceae (Gols & Harvey 2008). Olmez-Bayhan et al. (2007) have shown parasitism rates of the cabbage aphid are not reduced in higher-glucosinolate plants such as B. napus and Sinapis arvensis compared to B. oleracea, indicating that specialist parasitoids continue to provide some measure of pest control in high-glucosinolate systems. Whether or not other natural enemies, such as parasitoids, can compensate for lower syrphid predation is an open question and likely to vary among geographic locations and cropping systems.
15
A final consideration in determining the potential for B. nigra to serve as a source of aphids to surrounding crops is the production of winged morphs. Isothiocyanates, the volatile gas released in the mustard bomb upon predator attack, also cause release of the aphid alarm pheromone E-beta-farnesene (Bridges et al. 2002). The alarm pheromone triggers the production of more winged morphs in the population (Kunert et al. 2005), allowing aphids to escape high predator density to colonize another area with potentially fewer predators. From our study, it appears that glucosinolates may increase wing production even without predatory attack. Four out of the 31 aphids tracked to reproductive maturity on B. nigra developed into winged individuals, while none of the aphids on B. oleracea developed wings. As the aphid development experiment did not involve predators, wing production was presumably not driven by the release of isothiocyanates in this case; further study is necessary to determine if there are other mechanisms of glucosinolate-induced wing production. Increased wing production may augment the effect of increased total aphid production at the population level on B. nigra, facilitating the export of aphids to surrounding areas via increased flight. Field study documenting aphid movement to and from B. nigra would complement physiological research discerning the potential of this weed to serve as a source of aphids in an agricultural landscape. If aphids colonize weedy mustard at a higher rate than they do the crop, B. nigra may in fact be serving as an aphid sink. Tracking insect movement is difficult but has made advances in recent years (Hagler & Jackson 2001, Jones et al. 2006), and would greatly benefit our understanding of the impact of different Brassicas on aphid populations in the crop.
Conclusion
These results provide physiological evidence to suggest that B. nigra weeds surrounding a domesticated B. oleracea field can provide enemy-free space to cabbage aphids in agricultural landscapes without a physiological penalty. The lack of a significant reduction in aphid reproduction on B. nigra compared to B. oleracea suggests that there would be no long-term, population-level trade-offs of sequestering glucosinolates even when there is a nearly ten-fold difference in glucosinolate concentration in the adult aphid. Thus, the aphid has no reduced fecundity and a potentially increased survivability upon B. nigra, suggesting that B. nigra could serve as source of pest reintroduction to surrounding crops. The ability of aphids reared upon B. nigra to compromise the development and survival of their main predator indicates that this weedy mustard can provide a refuge from predation. While it remains possible that B. nigra provides only syrphid-free rather than enemy-free space, the existence of a refuge from this important predator should be a serious consideration for pest control.
The enhancement and promotion of natural enemies within the agricultural landscape has become the focus of much effort in pest control in recent years (Gurr et al. 1998). Their habitat requirements for food, mating, and overwintering is the main concern of conservation biological control. Maintaining healthy enemy communities is certainly important to the viability of biological control, but the issue of pest control must be 16 considered from the bottom up as well as the top down. Enemy-free space lifts these top- down constraints on pests, allowing them to build up and potentially spill over into other areas, overwhelming even functioning enemy communities. Habitat that may provide enemy-free space near crops should be identified and investigated for the threat it poses to agricultural production and security. Furthermore, studying the impact of enemy-free space on herbivores, their host plants, and their natural enemies helps illuminate the role phytochemistry plays in trophic ecology. Food web interactions in a heterogeneous environment such as the wild-crop interface are complex and often unpredictable, and enemy-free space presents an additional intricacy that must be acknowledged if we hope to better understand these systems.
17
Chapter 3: Pest Control Experiments Show Benefits of Complexity at Landscape and Local Scales
ABSTRACT: The ecosystem service of pest control provided to farms by natural enemies depends in large part on the setting in which those farms are found. While it is well established that natural habitat around farms enhances the natural enemy community on those farms, there is little or no evidence that such enhanced natural enemy communities elevate the level of pest control on the farm. This study measured the effect of habitat complexity on pest suppression of the cabbage aphid in broccoli, and reveals the importance of complexity at the landscape level as well as at the local on-farm level late in the growing season. Proportional reduction of pest densities (PRD) was four times higher at more complex sites. The degree to which complexity at one scale could compensate for a lack of complexity at the other was dependent on aphid colonization and population growth early in the season, determined by factors such as temperature and dispersal patterns. Managing habitat at different scales to promote natural enemy communities is an important consideration for pest control, but must be employed with other methods of pest management for truly effective and reliable control.
18 Introduction
Natural enemies, the predators and parasitoids of agricultural pests, could provide a sustainable and efficient alternative to pesticides. In order to capitalize on the ecosystem services provided by these biological agents of pest control, it is important to understand the conditions under which top-down control can and cannot maintain pest populations below desired levels. The challenge in characterizing the role of natural enemies in pest control is differentiating it from the many other variables influencing pest distributions. The magnitude of pest control provided by natural enemies is often only discovered upon enemy removal. The many documented cases of post-insecticidal rebounds and subsequent population explosions of pests underscore the general importance of a healthy and intact natural enemy community in reducing pest densities (reviewed by DeBach & Rosen 1991). The hundreds of success stories of pest control achieved through novel introductions of biological control agents provide further evidence of the capacity of natural enemies to constrain pest populations (Gurr & Wratten 2000). Dramatic pest reductions have also been recorded for communities of parasitoids and predators in experimental settings, often with the possibility of additive or greater than additive (facilitative or synergistic) effects (Snyder & Ives 2003, Schmidt et al. 2003, Losey & Denno 1998, Cardinale et al. 2003). However, the extent to which an enemy’s ability to reduce pest densities constitutes “control” is not entirely clear. The capacity for natural enemies to control pest populations in agricultural systems likely depends on local field conditions, not easily reproduced experimental settings. This study investigates some of the ecological and environmental factors that constrain or enhance the potential contribution of biological agents to the management of crop pest problems.
The study of ecosystem services provided by the natural enemies of pests is encompassed by conservation biological control, the goal of which is to maximize these services through the creation or improvement of habitat for native or at least naturally-occurring natural enemies. Such habitat management may include planting hedgerows or weed- strips along field edges, creating strips of unmanaged habitat within the field, leaving adjacent fields fallow, inter-planting crops to provide additional floral resources, and/or undertaking larger-scale efforts such as restoration or conservation of perennial habitat surrounding the farm (Fiedler et al. 2008). Habitat manipulation has met with mixed results at the local scale, and scientists are increasingly recognizing the importance of the landscape scale to maintaining natural enemy communities. Structurally complex landscapes, defined as landscapes with high proportions of natural or unmanaged habitat that are often also highly heterogeneous, are associated with increased abundance and diversity of natural enemies on farms, across many different geographic locations and crop types (reviewed by Kremen & Chaplin-Kramer 2007; Drapela et al. 2008, Gardiner et al. 2009b, Oberg et al. 2008, Schmidt et al. 2008, Werling & Gratton 2008). The resources provided for beneficial insects by complex landscapes may complement or substitute for local diversity in agro-ecosystems (Tscharntke et al. 2005). The reverse may not hold true, however. It has been suggested that local modifications may not support a viable natural enemy community on their own, but only serve to concentrate enemies already supported by natural habitat in the broader landscape surrounding the farm (Gurr et al. 1998). Both scales are likely important to the maintenance of healthy 19 enemy communities, but the interplay between them is not well understood, and the distinction between them is inconsistent. Some studies have investigated complexity at local and landscape scales simultaneously, but it has proven difficult to resolve the relative contribution of each factor (Bianchi et al. 2008, Clough et al. 2005, Elliott et al. 1999, Letourneau & Goldstein 2001, Östman et al. 2001b, Thies & Tscharntke 1999, Werling & Gratton 2008). This is likely because local and landscape complexity are often correlated; small diverse farms tend to be found in small wooded valleys and large monocultures are more likely to be in vast agricultural areas, making it difficult to differentiate between these two factors.
It is generally presumed in these studies that enhancement of enemy communities translates into a pest control benefit, based on previously established relationships between pests and natural enemies in laboratory or controlled field environments. However, studies that have considered the impact of landscape on pest densities have not found a consistent relationship between the two (Jonsen & Fahrig 1997, Kruess 2003, Letourneau & Goldstein 2001, Östman et al. 2001a, Roschewitz et al. 2005, Thies et al. 2005, Vollhardt et al. 2008, Zaller et al. 2008).
If landscape complexity enhances natural enemy communities, why is there no apparent relationship between landscape complexity and pest densities? The apparent disparity in the ability of natural enemies to reduce pest densities in experimental (e.g., Snyder & Ives 2003, etc.) versus agricultural (e.g., Jonsen & Fahrig 1997, etc.) settings may be due to endogenous variation in pest populations over space and time. Pests are often characterized by patchy distributions, and their densities would likely vary in different landscapes even without the influence of top-down control. Thus, the impact of natural enemies on pests may be substantial, but hidden by other mitigating factors; top-down control by natural enemies may be necessary but not sufficient to establish effective pest control. To more effectively elucidate the role of natural enemies in controlling pest populations, the importance of landscape complexity to natural enemies must be weighed against its impact on other factors affecting pest distributions (such as its potential to provide alternate host plants and/or predator refuges for pests; see Chapter 2).
The first step in understanding how landscape complexity contributes to overall pest management is to isolate the impact of landscape-mediated natural enemy increases on pest population dynamics. Pest densities alone are an inadequate measure of the extent to which pests are impacted by top-down control, since the underlying variation in aphid distributions (and thus what their densities would be in absence of control) is unknown. Cage experiments can be used to determine the amount of pest control provided by natural enemies, and employing such experiments over a landscape gradient addresses the question of whether increased natural enemy density or diversity in more complex landscapes actually confers a pest control benefit. Our hypotheses are that increased habitat complexity (i.e., complexity and both landscape and local scales, compared to one or the other) does provide superior pest control services to farms, and that one scale of complexity can substitute for another. The aims of this study are to explicitly measure the effect of complexity at different scales, to distinguish the role of natural enemies from other factors affecting pest populations, and to determine whether and under what
20 conditions the ecosystem service of pest control by natural enemies can contribute to a successful pest management system.
Materials and Methods
Study system The study was conducted over two years (2008-09) on 10 organic broccoli farms on California’s Central Coast, in Santa Cruz, Monterey, and San Benito Counties (Figure 3.1). Cabbage aphids (Brevicoryne brassicae) are a major pest of broccoli. Their natural enemies include Syrphid flies (Syrphidae), parasitic wasps (Diaeretiella rapae), lady beetles (Coccinellidae), lacewing larvae (Chrysopidae), spiders, and a variety of other coleopteran and hemipteran predators. Most of these enemies are extremely mobile and forage in many different habitats for floral resources and/or alternate prey. Syrphid larvae have been shown to be the most effective predators of aphids in this system (van Emden 1963). They require pollen as adults to build energy reserves for reproduction, making floral resources on farms or in the surrounding habitat an important determinant of their distribution in crop fields (Schneider 1969).
Study sites were selected at either end of a landscape complexity gradient, defined by the amount of natural habitat surrounding the farm, ranging from <10% to >50% of the total area within 2.5 km of the farm (Table 3.1). This scale was chosen based on a previous analysis of the relationship between abundance of the dominant predator (syrphid fly larvae) and landscape composition at multiple scales from 0.5 to 3 km. Natural habitat cover at a radius of 2.5 km was the best predictor of syrphid densities (r2 = 0.51, p < 0.001, Chapter 4). Sites with high proportions of natural or non-crop habitat in the surrounding landscape would typically be defined as “complex” (e.g., Vollhardt et al. 2008), but since complexity can refer to both the landscape and local scales, we will use different terminology for clarity. Sites with >50% natural habitat in the 2.5 km around the farm will be designated “natural matrix” landscapes and those with <10% natural habitat will be designated “agricultural matrix” landscapes, since the dominant land-use in these landscapes is agriculture. Four sites in each landscape category were selected for study in 2008, for a total of eight sites, and one additional site in each category was included in 2009, for a total of 10 sites. The study sites also differed in local habitat complexity. “Locally complex” farms were characterized by smaller fields, a diverse array of crops, and the incorporation of non-crop floral resources for beneficial insects in the form of hedgerows or weed strips along the field edge. “Locally simple” farms were large monocultures lacking any non-crop habitat within the farm itself. Two farms of each type of local category were selected for each landscape category in the 2008 study. In 2009, two “locally simple” farms and three “locally complex” farms were used in each landscape category (Table 3.1).
The landscape surrounding the farms was characterized using Geographic Information Systems (ArcGIS, Version 9.3.1, Environmental Systems Research Institute). Aerial photographs at 1 m resolution were obtained through the National Agricultural Imagery 21
22
23
Project (http://gif.berkeley.edu/resources/naip.html) for the area around each farm site. The photographs were digitized using an object-based image analysis program (eCognition, version 5.0, Definiens Inc, Alexandria VA), and the resulting maps were classified by hand to differentiate between agricultural and natural habitat. Proportional areas were then computed for each land-use class at a radius of 2.5 km around the farm site using Hawth’s tools (version 3.27, http://www.spatialecology.com).
Cage Experiment This study employed a similar approach to that used by Gardiner and colleagues (2009a), who demonstrated enhanced pest control benefits in more complex landscapes for soybean aphid through a cage experiment. However, their design involved placing cages over crop plants with naturally-occurring aphid densities, a method that would lose resolution at the low aphid densities resulting from low recruitment (as is seen in this system in some years, Chapter 4). With pests that are extremely patchily distributed, such as aphids, holding initial densities constant provides a better measure of pest control function over a landscape gradient. Comparing aphid growth in different locations with the same initial number of aphids measures the differences in pest control provided by resident natural enemies, independent of the natural rate of aphid settlement. Our study used cages containing plants inoculated with a pre-determined aphid density to achieve this comparison.
Broccoli plants were transplanted into pots from one-month-old starts (acquired from Growers Transplanting, Salinas, CA) in the greenhouse one month prior to the start of the experiment. These experimental plants were comparable in size to young broccoli plantings on a commercial farm, which generally take two to three months from transplanting to harvest. The potted plants were then set out in individual cages at each farm site for 12 days. The cages were constructed out of a fine organdy mesh to allow sunlight, precipitation and wind to pass through relatively unencumbered while excluding even the smallest natural enemies. Each cage sat on a plastic stage over a bucket of water, with cotton rope extending from the soil in the pot through the plastic stage into the water beneath it. Through a wicking action, the plants received sufficient access to water for the duration of the experiment. Data-loggers recorded temperature and humidity at 10-minute intervals in each of the treatments at each site to determine if there were any microclimate differences between treatments (Hygrochron I-buttons, Embedded Data Systems, Lawrenceburg, KY). No significant differences were found.
Aphid densities on broccoli crops one month after transplanting can range from zero to as many as a hundred aphids per plant, depending on season and location (Chaplin-Kramer, unpublished). Based on this observation, the potted broccoli plants used in this cage experiment were inoculated with 50 cabbage aphids each, well within the normal range of 0-100. Infested leaves from greenhouse cabbage-aphid colonies were placed on experimental plants after first removing any winged morphs, and the remaining aphids were given several days to transfer from cut leaf to living plant. Once the transfer was complete, aphids on the experimental plants were counted and aphids in excess of 50 were removed. Pilot studies had shown that cabbage aphids remain attached to the leaf 24 surface during transport to the field site; plants were re-checked in the field to confirm this at the start of each experiment.
The experiments were carried out in August 2008 and twice in 2009: in June, the early season before the aphid peak that occurs in mid-July (Chapter 4), and again in August, the late season after the aphid peak. Cages were arranged in replicate groups (three per site in 2008, two per site in 2009) at the edges of broccoli fields to standardize for differences in field size (cf. Kremen et al. 2004). Each replicate had one closed cage, two open cages, and one sentinel cage (defined below), for a total of 12 cages per site, 96 cages total in 2008; eight cages per site, 80 cages total in 2009.
The closed cages served as the control, measuring the population growth of aphids in absence of predation over 12 days at each site. The open cages served as the experimental treatment, measuring the population growth of aphids when exposed to predation and/or parasitism by natural enemies. The open cages had more replicates because this group showed the greatest variance in pilot studies, depending on whether and when natural enemies found the aphids on these plants. The sentinel cages were identical to the open cages, but no aphids were placed on these plants. These sentinel cages were thus a secondary type of control, to account for the possibility of locally- occurring aphids settling on the experimental plants, since the open cages would allow additional aphids as well as natural enemies to colonize the plants. Plants were harvested at the end of the 12 days and brought back to the lab. When the plants were harvested from the cages, leaf matter from 20 plants in the adjacent broccoli field was also collected for each site, to provide a comparison of aphid and syrphid larvae density (per gram of plant material) in the cages versus at the broader field level. Each plant or leaf was washed over a sieve to remove all the insects inhabiting it, and the insects were then identified and counted using a dissecting scope.
The experiment measured the ability of resident natural enemies to constrain aphid population growth, or the proportional reduction in aphid density in the open cages (corrected by the sentinel cages) as compared to the closed cages, which were free of control from predators and parasitoids. The proportional reduction in aphid density (PRD) was calculated as:
PRD = 1 – ((DO – DS) / DC), where DO is the average density of aphids on the plants in the open cages at the end of the experiment, DS is the average density of aphids in the sentinel cages, and DC is the average density of aphids in the closed cages. This measure is considered in terms of relative pest reduction, rather than pest control, since it does not necessarily correspond to lower absolute aphid densities at the overall field level (as discussed below). A caveat to these values for PRD is that they are more instructive to use as comparisons between sites than as absolute measures. For example, PRD “exceeding” 1 was an artifact of plants in the open cages having lower aphid densities than even the sentinel plants, a possibility when the natural rate of aphid settlement is high. Since aphid population growth is strongly dependent on temperature (Dixon 1977), climatic differences between
25 sites can result in vastly different intrinsic growth. Therefore, considering a relative measure such as PRD, instead of final aphid densities, allows a clearer comparison between landscapes because it controls for differences in aphid settlement and population growth rate at different sites.
The experimental design also allowed for additional comparisons between sites, including reproduction and colonization. Net reproductive rate (R) was measured as the change in aphid densities within the closed cages, or:
R = DC(T+1) /DC(T), where DC(T+1) is the average aphid density within the closed cages at the end of the experiment and DC(T) is the initial aphid density within all closed cages, 50 aphids. Colonization (C) was measured as the average aphid density on the sentinel plants at the end of the experiment (DS(T+1)). Combining the total number of aphids produced from reproduction (in early and late season) with total number of aphids arriving from colonization (in early and late season) provides an idea of the aphid pressure (P) that each farm faces:
P = (DC(T+1) - DC(T)) [early] + (DC(T+1) - DC(T)) [late] + (DS(T+1)) [early] +(DS(T+1)) [late]
Analysis Differences in PRD were assessed with a Multi-Factor Analysis of Variance using the statistical program R (version 2.9.1, http://www.R-project.org). Local (simple or complex) and landscape matrix (natural or agricultural) were factors for the 2008 analysis; local, landscape, and season (early or late) were employed as factors for the 2009 analysis. Local by landscape matrix interactions were included, as were interactions between each of these factors and seasonality. Pest pressure (P) was added as a covariate to determine if it improved the model. Pest pressure was also analyzed in an ANOVA against local, landscape, and seasonal factors. The natural log of net reproductive rate (Ln[R]) and colonization (C) were each analyzed with a linear regression against each other, temperature and/or the presence of weedy mustard, which may serve as a possible source of aphids to crop fields (Chapter 2).
Abundance of syrphid larvae within the cages and in the broader fields was log- transformed and compared to landscape and local factors using ANOVA. Syrphid abundance was also compared to the PRD found at each site with a linear regression, in order to determine whether complexity at either scale was linked to syrphid densities and whether higher syrphid densities resulted in greater pest control services. Aphid densities in the broader field were log-transformed and analyzed in a linear regression against PRD to test whether higher levels of relative pest reduction resulted in lower absolute pest densities. Field syrphid density, aphid colonization rate, aphid net reproductive rate, temperature, season, landscape and local complexity were added stepwise as covariates to determine if any of these factors improved the model.
26 Results
The 2008 experiment revealed that the PRD in natural matrix landscapes (>50% natural habitat) was significantly greater than that found in agricultural matrix landscapes (<10% natural habitat; Figure 3.2a, Table 3.2a). In addition to this landscape effect, significant local and interaction effects were detected in 2008 (Table 3.2a). Specifically, there was no effect of local complexity in the more natural matrix landscape, but in the agricultural matrix landscape, locally complex sites had nearly four times the PRD of locally simple sites (Tukey test for Honest Significant Differences, p < 0.001; Figure 3.2a). Likewise, there was no effect of landscape matrix at locally complex sites, but at locally simple sites, natural matrix landscapes showed a similar four-fold pest advantage over agricultural matrix landscapes (HSD, p <0.001).
The experiment was repeated twice in 2009 to determine if the same trend was seen in the early season as the late season, before and after the aphid peak. Again, there were significant local and landscape effects, as well as significant interactions between all factors (local complexity, landscape matrix type, and season; Table 3.2b). The late season trend mirrored that found in late-season 2008 (Figure 3.2b); PRD was enhanced at locally simple sites by a natural matrix landscape (HSD, p = 0.001) and in agricultural matrix landscapes by local complexity (HSD, p = 0.002). However, a substantially different trend was found in the early season. Landscape matrix type was the overriding factor in determining PRD, with sites in a natural matrix exhibiting five to ten times the proportional reduction of sites in an agricultural matrix (HSD, p < 0.001). Local factors did not play a role in the early season. The main difference between early and late season was found in the locally-complex agricultural matrix sites (HSD, p = 0.003). These sites reached nearly as high levels of PRD as those found in the natural matrix landscapes in the late season, but were indistinguishable from the simplest sites (locally-simple agricultural matrix) in the early season.
Pest pressure improved the model for PRD, though its relative contribution was weak compared to local, landscape and seasonal factors (partial r2 =0.01, p < 0.001). Pest pressure itself was also impacted by these factors, increasing with landscape and local complexity (p = 0.001, Figure 3.3, Table 3.3). This may have been partly due to a weak trend of higher maximum temperatures in natural landscapes in the late season (ANOVA, F = 4.77, df = 1,8, p = 0.06). Higher maximum temperatures drove higher net reproductive rates (partial r2 = 0.27, p < 0.001; Figure 3.4). Net reproductive rate, in turn, drove colonization, along with the presence of weedy mustard around farm sites (r2 = 0.84, p < 0.001; Figure 3.5).
Abundance of syrphid larvae found in the cages was correlated with that found in the fields (r = 0.77, p < 0.01), which were an order of magnitude higher in natural than agricultural matrix landscapes (ANOVA, F= 12.44, d.f. = 1,8, p = 0.007). Higher densities of syrphid larvae in the fields were associated with higher PRD (r2 = 0.49, p = 0.023). Field aphid densities showed a weak trend of declining with PRD when net reproductive rate was also included in the model (Table 3.4). No other factors improved the model, but there was a significant seasonal effect on aphid densities in a separate
27
28
29
30
31
32
33 analysis (mean field aphid density across all sites in the early season: 25 aphids / 100g plant material; late season: 449 aphids/ 100g plant; F = 11.31, DF = 1, 18, p = 0.003, ANOVA on log-transformed data).
Discussion
This study demonstrates that habitat complexity improves pest control provided by natural enemies on farms. Complexity at the local scale (crop diversity, floral resources) can substitute for that at the landscape scale (natural habitat) or vice versa, late in the season after the aphid peak. Sites with high local complexity and low levels of natural habitat in the landscape matrix had the same level of PRD as sites with low local complexity and high levels of natural habitat in the landscape matrix in the late season. The PRD found at these mixed complexity sites (locally-complex agricultural matrix and locally-simple natural matrix) was only slightly lower than at the most complex sites (locally-complex natural matrix landscapes), but substantially higher than the simplest sites (locally-simple agricultural matrix landscapes). Not only the direction of this trend, but the magnitude was similar between years, with the mixed complexity sites exhibiting an 80% pest reduction, four times greater than the simplest sites and not significantly different from the sites that were complex at both the local and landscape levels. This suggests that local complexity is more valuable for natural enemies in simpler (agricultural) landscapes than in complex ones, corroborating the findings of Haenke et al. (2009).
The apparent substitutability of complexity across scales in the late season was not evident in the early season. The low levels of PRD found in the locally-complex agricultural matrix landscapes in the early season indicate that there may be a lag-time in the response of natural enemies at these sites. While the locally-complex agricultural landscapes then “caught up” to the natural landscapes later in the season, this early- season lag in pest reduction is an important consideration, with implications for the overall ability of enemies to constrain aphid populations at the field scale. Before the aphid peak, aphid densities in the fields were orders of magnitude lower than after the peak. If enemies can better constrain aphids at their low levels early in the season, the aphid peak may subsequently be less pronounced. However, a lag in enemy response, as seen in the locally-complex agricultural matrix landscapes, may create a window of opportunity for rapid aphid population growth.
Previous work has demonstrated that predation during the early stages of aphid establishment determines total population and ultimately crop yields more so than predation later in the season (Östman et al. 2001a). Therefore “pest control” is context- dependent, and the high levels of PRD seen in the locally complex agricultural landscapes later in the season may be too little, too late. Indeed, a three-year insect survey (Chaplin-Kramer, unpublished) showed that the locally complex agricultural matrix sites did not reach syrphid densities of two per plant until two weeks or less before harvest, while such densities were already found at many of the natural matrix sites 34 (including some of the locally simple ones) by the first week of sampling, five to six weeks before harvest. These locally complex agricultural matrix sites may offer enough habitat at the farm level to attract natural enemies but not enough within the surrounding landscape to sustain them. Enemies arriving on the scene later in the season are likely drawn to the substantial prey resource in the fields, but may just be skimming off the top, rather than providing true top-down control.
Aphid distributions are obviously determined by many factors other than top-down control. These additional factors must be considered in order to truly understand and anticipate pest pressure and the contribution natural enemies can make to effective pest control in agricultural settings. Warmer temperatures that enhance aphid reproductive rates could potentially overwhelm an enemy community that might otherwise provide effective control in cooler conditions. Given the (albeit weak) relationship between temperature and landscape complexity, it is possible that the lack of vegetative structural diversity in the simplest landscapes directly impacts the microclimate at those sites.
Another factor determining aphid distributions is colonization, presumably influenced by prevailing wind patterns and the presence of alternate host plants, since aphids are not strong fliers (Dixon & Howard 1986). In areas of high dispersal, a constant supply of aphids replenishing those lost to predation may mask the pest control service provided by natural enemies. Our results indicate that the natural matrix landscapes had higher aphid colonization rates in the late season, potentially driven in part by the presence of weedy mustard, which indicates the importance of the specific composition of natural habitat. The presence of alternative host plants raises a broader question about how the distribution of weeds can impact the delivery of ecosystem services. Some types of natural habitat harbor large patches of weeds, which often include host plants for pests (Landis et al. 2000). This is especially true in the heavily invaded annual grasslands found around the UCSC site, the study location with the highest aphid colonization. These invaded grasslands were designated as “natural” for the purposes of this study, but a distinction could be made between truly natural and simply unmanaged habitats. In contrast, the two locally simple agricultural matrix sites (CPS and TAS) are characterized by an almost complete lack of non-crop habitat, proactively managed for removal of weeds like mustard along the field edges. These sites exhibited notably low aphid colonization rates. Future work should examine the role of natural or unmanaged habitats in supporting not only natural enemies of pests, but also the pests themselves.
Although habitat complexity may promote pest population growth by influencing microclimate or providing resources to pests, our results clearly indicate that natural habitat near farmland provides pest control services to farms. Sites in the natural matrix landscapes, even with higher aphid pressure, exhibited higher pest control than sites in the agricultural matrix landscapes. Including factors such as colonization and net reproductive rate did not improve the model for PRD, suggesting that pest control can supersede high levels of colonization and high growth rates, at least at the low initial densities found in the cages. At higher aphid densities, such as those found in the crop later in the season, PRD was a less predictive factor in determining field aphid densities than net reproductive rate. The presence of a healthy natural enemy community on the
35 farm is therefore even more important in the early season, when aphid populations are small and not increasing as rapidly, and when natural enemies have the best chance at keeping populations from growing.
Conclusion
This study demonstrates that habitat complexity is an important consideration for pest control services; proportional reduction of pest density increases with natural habitat at the landscape scale. While high local complexity can compensate for low landscape diversity and vice versa, a lag-time in natural enemy arrival in simpler (agricultural matrix) landscapes may compromise the enemies’ ability to provide adequate control, depending on site-specific conditions such as temperature or dispersal patterns. Local complexity in the absence of landscape complexity may provide adequate top-down pest control in cooler microclimates with relatively low aphid colonization. However, even exceptional natural enemy function can be overwhelmed by high enough rates of pest reproduction or colonization from nearby source habitat.
While the landscape scale appears to be more important to pest reduction than the local scale, growers typically have more control over their local habitat than the surrounding landscape. It is important to recognize that good on-farm management, especially the creation of habitat for natural enemies via hedgerows or insectary-strips and the exclusion of alternate host plants for the pests, can benefit pest control in certain circumstances. The collective local actions of many growers could even translate into a broader landscape effect, as has been shown for pollinators (Holzschuh et al. 2008). However, such management may not be sufficient to constrain pest populations below economic thresholds, and farmers may need to employ other methods to achieve the desired level of control (Zehnder et al. 2007). Other aspects of local complexity, such as trap-cropping or intercropping, may contribute to pest reduction through bottom-up means as well, as described by the Resource Concentration Hypothesis (Root 1973, Andow 1990).
Likewise, complexity at the landscape scale is not a panacea for pest problems. Efforts to restore or create larger-scale habitat for beneficial insects must consider which types of habitat to promote. Habitat restoration is often the purview of local government or non- profit groups whose interests may focus more on other aspects of conservation than on the impact of habitat on pest control services. However, it may be possible to maximize habitat for beneficial insects while minimizing its potential to serve as a refuge for pests. The major challenge for pest management is identifying such pest refuges and understanding their contribution to on-farm pest problems.
The important message for growers and land managers is that habitat complexity improves pest control services – at all scales – but may not be the only part of the equation that matters. Pest infestations on diverse farms and/or in natural landscapes do not necessarily indicate a failure of complexity to enhance natural enemy suppression of pests; the pest problem could be much worse without it. However, pest suppression is 36 not the same as pest control. In many cases, it will be necessary to consider other variables contributing to pest distributions in addition to top-down factors in order to achieve truly effective natural pest control.
37 Chapter 4: Subtle Effects of Pest Control Services Along a Landscape Gradient and Across Temporal Scales
ABSTRACT: Natural habitat may deliver ecosystem services to agriculture through the provision of natural enemies of agricultural pests. Many studies have documented positive effects of landscape complexity on natural enemies but fewer have examined the simultaneous resulting impact on pests. This study measured the effect of landscape composition on different indicators of pest control services in a three-year study in Central California broccoli. The results showed that abundance of natural predators (syrphid fly larvae) increased strongly with the proportion of natural habitat surrounding the farm, cabbage aphid population growth was reduced on farms with higher syrphid densities, and final aphid densities (the week before harvest) declined with natural habitat. Pest control services delivered by natural habitat may be masked by inter-annual variation in environmental factors and by competing direct and indirect effects (i.e., landscape effects on the pests themselves versus on the natural enemies’ control of pests).
38 Introduction
Predators and parasitoid insects can provide a valuable ecosystem service to agriculture through reduction or control of pest populations. It has been estimated that current crop losses to pests represent only about 35% of losses that could be expected in the absence of this ecosystem service, which translates to an annual value of natural enemies that exceeds $13 billion in the U.S. alone (Losey & Vaughan 2006). Better management of agricultural landscapes to enhance natural enemy populations could greatly increase their potential value. Ecosystem services provided by natural enemies differ fundamentally from those produced in situ such as nutrient cycling or carbon storage because they are delivered by mobile organisms (Kremen et al. 2007). Pest control services to farms are provided by natural enemies that may have foraged in habitat far beyond the farm’s boundaries in order to obtain their needed resources. Understanding the delivery of pest control services therefore requires a landscape-scale perspective (Tscharntke et al. 2007). However, of the many landscape-scale studies that have measured natural enemy abundance or diversity as a metric of “pest control”, few have documented how the landscape impacts pest distributions and their impact on crops. This study examines the role of natural habitat in providing enemies to perform pest control services. Weekly insect surveys over three years explore the effects of landscape composition on natural enemies (syrphid larval abundance), pests (aphid population growth), and in turn the provision of the pest control service (reduction of infestation of broccoli).
Landscape complexity, often measured by landscape composition or diversity, has been widely shown to benefit populations of natural enemies of agricultural pests (reviewed in Bianchi et al. 2006, Kremen & Chaplin-Kramer 2007). Among a vast array of regions and crop systems, predators and parasitoids increase in abundance (Clough et al. 2005, Elliott et al. 2002, Gardiner et al. 2009a, Gardiner et al. 2009b, Letourneau & Goldstein 2001, Prasifka et al. 2004, Schmidt et al. 2005, Schmidt & Tscharntke 2005, Schmidt et al. 2008) and diversity (Clough et al. 2005, Drapela et al. 2008, Purtauf et al. 2005, Schmidt et al. 2005, Schmidt et al. 2008, Steffan-Dewenter 2002, Werling & Gratton 2008) with the proportion of non-crop (unmanaged, natural or semi-natural) habitat in the surrounding landscape. Non-crop habitat can also be associated with a shift in enemy community assemblage (Elliott et al. 2002, French et al. 2001, Gardiner et al. 2009b, Isia et al. 2006), and such structural changes could have functional consequences since the different enemy species can interact synergistically (Cardinale et al. 2003). While the majority of studies demonstrate a positive relationship between natural enemies and non- crop habitat, some have found that enemy responses may vary over time (Menalled et al. 2003, Oberg et al. 2008, Prasifka et al. 2004, Schmidt & Tscharntke 2005) and space (Menalled et al. 1999) or between species (Menalled et al. 2003, Schmidt et al. 2008, Zaller et al. 2009). However, confounding environmental factors such as climate (Schmidt & Tscharntke 2005, Zaller et al. 2009) and host plant presence (Menalled et al. 1999, Menalled et al. 2003) may provide alternative explanations for these apparent exceptions.
Increases in enemy abundance and diversity are often associated with higher levels of pest mortality by natural enemy action, and similar positive relationships between non- 39 crop habitat and rates of parasitism or predation have been documented in many systems (Bianchi et al. 2008, Bianchi et al. 2005, Boccaccio & Petacchi 2009, Gardiner et al. 2009a, Steffan-Dewenter 2002, Thies & Tscharntke 1999, Thies et al. 2008, Thies et al. 2005, Tscharntke et al. 2002). A few studies have revealed cascading effects of landscape complexity increasing parasitism and reducing plant damage by pests (Thies et al. 2003, Thies & Tscharntke 1999). However, increased parasitism or predation does not always imply a reduction in pest densities or plant damage. In fact, parasitism is positively correlated with pest densities in some systems, suggesting that pest densities sometimes drive parasitoid activity rather than vice versa (Thies et al. 2005, Costamagna et al. 2004). Further, increased parasitism can be offset by greater pest colonization in more complex landscapes, resulting in no net change in pest densities over a landscape gradient (Roschewitz et al. 2005, Thies et al. 2005). For these reasons, parasitism rate alone is not an adequate measure of pest control.
In order to truly assess landscape effects on pest control, research must move beyond documenting only enemy response to documenting the resultant pest response as well. A review of 10 such studies found that only half reported a negative association between pests and landscape complexity (Bianchi et al. 2006). These include the two previously mentioned studies demonstrating reduced crop damage with increased parasitism (Thies et al. 2003, Thies & Tscharntke 1999), but the other three are less conclusive. Östman et al. (2001a) saw lower aphid establishment but also documented greater population growth later in the season in more complex landscapes. Den Belder et al. (2002) reported reduced pest densities in leek fields near woodlots, but indicated this was not the action of natural enemies, since few were found in this system. Rather, den Belder et al. suggested that in this case landscape complexity reduced pest densities by inhibiting dispersal. The final study Bianchi et al. (2006) presented as evidence for a negative relationship between landscape complexity and pest density was of questionable relevance to the “landscape” scale, because it focused only on field boundaries and hedges (Basedow 1990).
Several studies not included by Bianchi et al. (2006) further confound the relationship between landscape complexity and pest densities. Zaller et al. (2008) found herbivore abundance increased with landscape complexity, in the same system in which Thies and colleagues (Thies et al. 2003, Thies & Tscharntke 1999) documented a decrease, although the former recognized that parasitoids were rare in their study region and would therefore not provide the level of control documented by the latter. Other studies have found positive relationships with landscape complexity for some species and negative or neutral relationships for others (Jonsen & Fahrig 1997, Letourneau & Goldstein 2001). To complicate matters further, there may be non-linear relationships between crop area, non-crop area, and pest distributions. Valantin-Morrison et al. (2007) found that a threshold existed for crop area, below which, landscape complexity was negatively correlated with pest densities, but above which, complexity was positively correlated with pest densities.
Some of these studies may have failed to find an impact of landscape on pest control because they were not measuring it correctly. Pest densities are driven by many factors
40 that may vary from site to site besides natural enemies (e.g., climate, host plant presence, dispersal patterns, etc.). Simply examining pest densities permits no inference as to how much more abundant pests would be in the absence of their natural enemies, if all other things were equal. Comparing instead proportional change in pest populations over time, with respect to natural enemies and/or landscape, may provide more insight into this question of how landscape complexity affects the ecosystem service of pest control. Furthermore, many previous studies investigating landscape effects have taken only a snapshot of pest and enemy densities at one point during the season or have limited their study period to one year. Studies in which samples were collected multiple times over the season and/or over multiple years have emphasized the importance of seasonal and inter-annual variation in their results (Menalled et al. 2003, Oberg et al. 2008, Östman et al. 2001a, Prasifka et al. 2004, Schmidt & Tscharntke 2005). Therefore, this research investigates the role of landscape complexity on weekly changes in densities of cabbage aphids (Brevicoryne brassicae) and their syrphid predators (Diptera: Syrphidae) in broccoli over three years. The questions addressed in this study are:
1) Does natural habitat near farmland enhance natural enemy populations on farms, and at which scales?
2) Do landscape-mediated increases in natural enemies reduce aphid population growth in crops?
3) Does natural habitat near farmland result in lower aphid densities before harvest?
Materials and Methods
Study system This research was conducted in California’s Central Coast region, where nearly half of the nation’s broccoli is grown. Cabbage aphids (Brevicoryne brassicae) are a major pest of broccoli and are prey for a number of natural enemies, including parasitic wasps (Diaeretiella rapae), lady beetles (Coccinellidae), lacewing larvae (Chrysopidae), spiders, and a variety of other coleopteran and hemipteran predators. However, the larvae of flies in the family Syrphidae are the most abundant aphid predator by far in this system (70% of all natural enemies were syrphids, across all sites in all three years; unpublished data). It has been suggested that syrphids may in fact be the most efficient predators of the cabbage aphid (van Emden 1963). Adult aphidophagous syrphid flies are extremely mobile, searching in many different habitats for nectar and pollen on which to feed, in order to obtain the energy necessary for reproduction (Schneider 1969). Many modern farms do not provide sufficient floral resources to meet the energy demands of syrphids, making the surrounding habitat an important variable in syrphid distribution on farms. Syrphid flies are therefore an excellent study species to use to explore the effects of landscape on pest control.
41 The study was conducted on 24 organic broccoli farms in Santa Cruz, Monterey, and San Benito Counties (Figure 4.1, Table 4.1) during the summer growing seasons of 2006- 2008. Due to the annual rotation schedules of these farms, different fields were used each year, and new sites were acquired if certain producers chose not to grow broccoli in a given year. Ten farms were sampled in all three years, 14 in at least two years, for a total of 15 farms in 2006, 16 in 2007, and 17 in 2008. The landscapes in which the farm sites were set spanned a complexity gradient ranging from less than 5% (simple) to more than 80% (complex) natural habitat within 1, 2, and 3 km radii of the farm (Table 4.1). Natural habitat was composed predominantly of grassland, shrubland, and woodland. Most of the farms were (locally simple) large-scale industrial organic operations, similar in most respects to conventional farms except for the restrictions on use of pesticides and synthetic fertilizers. The few (locally complex) farms that grew a more diverse array of vegetables in smaller quantities occurred on both ends of the landscape gradient (including Earthbound’s Main Farm in Carmel Valley and UC Santa Cruz’s Center for Agroecology and Sustainable Food Systems at the complex end of the landscape gradient, and the USDA experimental farms and Agricultural Land-Based Training Association in Salinas at the simple end).
GIS land-use mapping Geographic information (point data) was taken at the center of each sampling transect each year using a Global Positioning System (Trimble Geo XT), and imported into Geographic Information Systems (ArcGIS, Version 9.3.1, Environmental Systems Research Institute). Aerial photographs of 1m resolution were obtained through the National Agricultural Imagery Project (http://gif.berkeley.edu/resources/naip.html) for the 3 km area surrounding each farm site in each year. The photographs were digitized using an object-based image analysis program called eCognition (version 5.0, Definiens Inc, Alexandria VA), and the resulting maps were classified by hand into the following land-use categories: annual agriculture, perennial agriculture, fallow land, industrial, residential, road, bare, water, and natural habitat. Proportional areas were then computed for each land-use class at a radius of 0.5, 1, 1.5, 2, 2.5, and 3 km around the farm site using Hawth’s Tools (version 3.27, http://www.spatialecology.com). These proportional areas were used in univariate regression analysis (R, version 2.9.1, http://www.R- project.org) of syrphid abundance data to determine which scale was most predictive of syrphid populations (cf. Ricketts et al. 2001). The most predictive scale was then used to analyze all data.
Insect collection Ten broccoli plants per week were collected at each field site starting four to seven weeks before harvest and continuing every week until harvest. The variation in total weeks sampled occurred because the crops occasionally reached maturity more quickly or slowly than the grower’s anticipated schedule. Further, different growers had different planting schedules, and as a result the first week of sampling occurred anywhere between May and August across the study sites.
42
43
44
45
There was significant seasonal and interannual variation in insect abundance, as aphid densities across all sites peaked around July in 2007 and 2008 (p < 0.001; Figure 4.2). The variable planting dates were unavoidable due to studying many different farms in different locations, however, and there was no significant correlation between sampling date and farm placement within the landscape gradient.
Individual plants and the insects they harbored were collected every 5-10 meters along a transect starting from the edge of a field, to standardize edge effects in different-sized fields (cf. Kremen et al. 2004). Entire plants were sampled by cutting at the base, were then immediately placed in individual sealed plastic bags and were processed within 48 hours of return to the lab. Each plant was washed over a sieve, and all insects collected were stored in alcohol, to be later identified and counted under the microscope. This method is an important collection technique for syrphids to ensure that all individuals inhabiting a plant were detected. Syrphid larvae are primarily nocturnal, hiding in the folds of leaves during the day (Rotheray 1986), and may therefore be underestimated by visual observation in the field. Different species of syrphid larvae are difficult to distinguish as larvae, so those collected in 2007 were reared through pupation to determine the general composition of the community. The most common species was Allograpta obliqua, but other species reared included Eupeodes americanus, Eupeodes volucris, Syrphus opinator, Scaeva pyrastri, Sphaerophoria sulphuripes, Toxomerus occidentalis, and Platycheirus stegnus. Upon rearing it became evident that syrphid eggs could not be reliably differentiated from leafminer eggs (Liriomyza spp., Diptera: Agromyzidae), so only larval counts were used in these analyses.
Analysis Insect counts were log-transformed and analyzed using the statistical program R (version 2.9.1, http://www.R-project.org). These data were first analyzed for edge effects along the transects; since none were detected, the counts on individual plants were averaged over the ten plants collected at each site in each week. Syrphid and aphid data were analyzed using a general metric (e.g., abundance) for each site in each year (site-level data), or based on population changes from week to week (week-level data).
The method of collection used in this study only detected insects present on the plants themselves. Syrphids generally only oviposit on aphid colonies (Schneider 1969), so even if syrphids were present at a site they would not be found on a plant with very low aphid densities. Therefore, it was important for those analyses involving syrphids to correct for low aphid densities. We accomplished this in one of two ways: 1) sites whose densities never exceeded 100 aphids/plant (a lower bound for syrphid oviposition preference determined by (Chandler 1968)) at some point during the growing season were excluded from any analysis involving syrphids (eliminating 17 of 48 sites across all three years, 11 of which were from 2006, an anomalously low aphid year), or 2) maximum aphid values could be incorporated into the model as an additional covariate. The latter, as a site-level variable, would not be appropriate in the week-level analyses; however. When appropriate, both of these approaches are presented. 46
Abundance of syrphid larvae was averaged over all weeks for each site in each year for use in a univariate linear regression with proportion of natural habitat at six different scales (0.5 – 3 km) around the farm in order to determine which scale to use in all subsequent analyses. After the appropriate scale was identified, syrphid larvae or aphids were analyzed using generalized linear models with several possible predictor variables: proportional natural habitat, local site complexity (categorically defined as simple or complex, as above), year (tested as a categorical factor, since there were only three years), season (categorically defined as early or peak, based on aphid densities pooled across all sites over the range of sampling dates), relative week (until harvest, for the week-level data), the previous week’s aphid or syrphid density (for week-level data), and maximum aphid density (as defined above, site-level).
Three response variables related to aphid densities were defined: 1) Change in aphid density from week to week (∆ aphids), 2) Minimum Value of ∆ aphids (MV∆), and 3) Final Aphid Density. Change in aphid density from week to week (∆ aphids), was measured as:
∆ aphids = Ln (aphid density in week (t+1)/aphid density in week (t))
The measure of ∆ aphids was initially analyzed in a linear mixed-effects model with site as a random effect, since multiple points of data from the same site cannot be considered independent measurements. However, the standard deviation of the random effect was essentially zero (<0.001), so the random effect was dropped and a linear model was employed instead, with previous week’s syrphid density, proportion of natural habitat, relative week, season, and year as fixed variables. The Minimum Value of ∆ aphids for each site in a given year would correspond to either the lowest increase or the greatest decrease, depending on whether the aphid population was continually growing or if it peaked and declined at some point over the sampling period. Syrphids could be seen as constraining aphid population growth to some degree if the MV∆ was correlated with the density of syrphid larvae in the earlier week (t). MV∆ was therefore analyzed in a linear model including previous week’s syrphid density, proportional natural habitat, and year. Final Aphid Densities (the week before harvest) were also analyzed with a linear model including proportional natural habitat and year.
Results
The landscape scale of 2.5 km accounted for the greatest proportion of total variance in syrphid larval abundance (r2 = 0.30, df = 1, 46, p < 0.001, Table 4.2), although all scales tested were significant (p < 0.05). When sites with consistently very low aphid densities (<100/plant) were excluded, the relationship between syrphids and natural habitat was stronger (r2 = 0.51, p < 0.001, Figure 4.3a), but the most predictive scale remained consistent. Incorporating maximum aphid densities as a covariate (the second method for incorporating the effect of low aphid densities), also improved the model (r2 = 0.46, p < 0.001). Year, season, and local complexity did not significantly affect syrphid abundance (Figure 4.3b). 47
48
49
The measure of ∆ aphids decreased with the previous week’s syrphid larvae density (Figure 4.4a, r2 = 0.16, p < 0.001), meaning that greater declines or lower increases in aphid populations were associated with greater syrphid abundance in the previous week. The effect of syrphids on ∆ aphids was magnified when considering only sites with aphid densities exceeding 100/plant (Figure 4.4b, r2 = 0.32, p < 0.001). No other variable (week, year, season, local complexity or natural habitat) improved the model.
Syrphid larvae density accounted for more of the variance in the minimum value of ∆ aphids than in all ∆ aphids (Figure 4.5a, r2 = 0.26, Table 4.3). Again, a negative relationship was seen between minimum ∆ aphids and syrphid larvae in the week prior to the occurrence of minimum ∆ aphids. Proportion of natural habitat in the landscape did not improve the model, but when included in place of syrphids it performed nearly as well (r2 = 0.25, Table 4.3). Excluding sites with low (<100) aphid densities strengthened the relationship between minimum ∆ aphids and syrphids, and introduced significant interaction effects (Figure 4.5b, r2 = 0.69, Table 4.3). Including all sites but incorporating maximum aphid densities yielded a similar but slightly weaker result (r2 = 0.53, Table 4.3). Local complexity and season were not significant effects.
There was no significant effect of natural habitat on final aphid densities prior to harvest if each year was analyzed separately, but pooling these data revealed a significant effect. Aphid densities the week before harvest were lower in more natural landscapes across all three years (r2 = 0.39, p < 0.001). Local complexity, while not a truly significant effect itself (p = 0.052), improved the model and therefore was included ( r2 = 0.45, Table 4.4). The partial r2 for natural habitat was weak (0.05), meaning that it explained relatively little of the overall variation, but the slope of the negative relationship between aphids and natural habitat was significant (p = 0.031). Year was the dominant factor in the model because 2006 had much lower aphid densities compared to 2007 and 2008 (Figure 4.2, p < 0.001). Interactions between year, landscape, and local variables were not significant. When 2006 was removed, the year effect was no longer significant. Final aphid densities were more tightly correlated with local and landscape complexity (r2 = 0.30, p < 0.014), and the slope of the negative relationship with landscape complexity was much more pronounced (Table 4.4).
Discussion
Three years of weekly insect survey data in this study provide evidence of pest control services to agriculture provided by natural habitat. Week to week changes in aphid densities showed lower increases and greater declines associated with greater densities of syrphids, suggesting that aphid population growth was better suppressed by the more abundant predators found in more complex landscapes. That the maximum potential for aphid suppression, as measured by the minimum value for ∆ aphids, was nearly as well correlated with landscape as with syrphid densities provides supporting evidence that 50 51
52
53
54 natural habitat is an important factor contributing to pest control. However, natural enemy suppression of pest populations does not necessarily imply adequate pest control, which is determined by economic thresholds and not ecological criteria. Natural enemies may suppress pest population growth without reducing pest densities (Thies et al. 2005), or they may indeed reduce pest densities but not reduce them sufficiently to avoid plant damage. In either case, growers may find it hard to rely upon natural enemies as a sufficient means of pest control. It is nonetheless important to recognize the contribution that natural enemies make to an overall pest management program and to identify conditions that could increase that contribution.
In this study, aphid suppression by syrphids did lead to reduced aphid densities in the week before harvest, when the pest loads matter most to growers. This effect was not evident in any individual year, but when pooled across years, the data revealed a significant effect. While natural habitat explained a rather small amount of the total variance in aphid densities relative to year, the magnitude of the effect (its slope coefficient) was comparable to the magnitude of the year effect. The impact of removing 2006 from the model further emphasizes the importance of multiple years of study (Menalled et al. 2003, Prasifka et al. 2004, Schmidt & Tscharntke 2005), and suggests that a lack of relationship between landscape and pest distributions in any single-year study may not be an accurate representation of the true dynamics of the system. Understanding the factors behind such temporal variation is a major challenge. Aphid densities were orders of magnitude higher across all sites in 2007 and 2008 than in 2006, a year with an unusually long wet season (heavy rains through May). Rain prevents winged aphids from dispersing, and increased rainfall is associated with delayed aphid establishment (Thackray et al. 2004). Rain can also impact aphid survival, knocking them off plants, leading to mortality from impact or starvation (Hughes 1963). The late rains in 2006 may have contributed to delayed aphid establishment in the crops and ultimately prevented aphid populations from ever reaching their peak (Dixon 1977). Indeed, 2007 and 2008 populations show a distinct peak around mid to late July, while aphids in 2006 showed no real trend with season.
Another problem in detecting an effect of landscape complexity on pest density may be that it is an indirect effect several times removed—landscape impacts enemies which impact pest mortality, which impacts pest population growth, which impacts pest densities—and with each relationship more error is introduced. Adding to this problem and introducing perhaps even more uncertainty are the numerous potential direct effects of landscape on pests. Indirect effects (landscape on natural enemies on pests) and direct effects (landscape on the pests themselves) may be working in concert with or in opposition to one another, masking the true role of landscape in providing pest control services to farms via natural enemies.
Landscape may directly impact pests in conflicting ways. On the one hand, natural habitat embedded in agricultural landscapes may act as barriers to dispersal as pests attempt to move from one crop field to another, either by disrupting wind patterns (Lewis 1965, Pasek 1988) or by forcing pests to pass through a gauntlet of higher enemy
55 densities within the natural habitat matrix (Schmidt & Tscharntke 2005). Even if it did not disrupt dispersal, natural habitat could thwart colonization by reducing the resource concentration of monoculture cropping (Root 1973), making it more difficult for pests to find their hosts. On the other hand, the same resources provided by natural habitat to parasitoids and predators could also benefit some pests, especially lepidopteran pests that seek nectar as adults (Winkler et al. 2009). Natural habitat could further serve as a perennial refuge for pests, from which to continually recolonize annual crop fields, by providing alternate host plants or overwintering habitat (Pasek 1988). In our system, the climate is Mediterranean and thus aphids do not need to overwinter (Hughes 1963), but they may build up populations on the weedy mustard that commonly occurs in unmanaged habitats near agricultural fields (Chapter 2). If natural habitat could be assessed in terms of its potential benefit to natural enemies versus pests, a clearer relationship between landscape and pest control might emerge. While the extent of on- the-ground plant surveys that would currently be required to undertake such work make it untenable, the rate of improvement in geospatial databases may make this a viable option in the future.
Remarkably, despite these multiple interacting effects causing aphid densities to fluctuate dramatically from year to year, syrphid densities were influenced more by landscape composition than inter-annual variation. Furthermore, the range in scales of syrphid response to landscape was quite broad; significant at all scales measured (from 0.5 to 3 km), but most predictive at one of the largest scales (2.5 km). This is in contrast to Haenke et al. (2009), who found syrphids responded only to the smallest spatial scales (0.5 – 1 km), surprising given the high mobility of syrphid flies. Sarthou et al. (2005) showed a response of syrphids to landscape features at 2 km, the largest scale those authors measured. It is possible that through their excellent vision and strong flying, syrphids can respond to resource opportunities at many different scales (Schneider 1969). Spiders have also displayed a significant response across a wide range of scales (Drapela et al. 2008, Schmidt et al. 2008), in contrast to the narrower range parasitoids typically display (Thies et al. 2003, Thies & Tscharntke 1999, Thies et al. 2005). Such a contrast in sensitivity to scale in two types of natural enemies may indicate an underlying difference between the scales in which generalist and specialist natural enemies operate. Though syrphids are not true generalists, most syrphid species can feed on a diversity of aphid species. This breadth of diet may buffer them against environmental stochasticity and prey fluctuations. The robustness of the relationship between landscape complexity and syrphid abundance over time and across scales suggests that they may be a fairly reliable provider of pest control services. Elmqvist et al. (2003) reviewed several cases in which replication of ecological functions across a range of scales led to greater resilience of the ecosystem service in the face of disturbance.
Conclusion
A tiered set of questions could help to evaluate the magnitude of pest control service provided by natural habitat near farmland, akin to that proposed by Heimpel and Jervis (2005) in the narrower context of floral subsidies for parasitoids. When evaluating the 56 contribution of natural habitat to overall pest control, we should consider the following, in order of increasing degree of evidence: Are natural enemies enhanced? Are pest mortality rates increased? Is pest population growth slowed or even in decline? Are pest densities reduced? And finally, is plant damage reduced below the economic injury level? The most straight-forward scenario would be finding evidence for affirmative answers to all of these questions; the ecosystem service could be confidently measured and managed. However, any positive answers along the spectrum of questions provide evidence for some amount of pest control and a sense of how important landscape factors might be in achieving such control. In this study, natural enemies were enhanced, pest population growth was slowed, and pest densities were weakly but statistically significantly reduced. Plant damage by aphids is generally not considered a threat in broccoli. The greater concern is aphid infestation of the head, since it is nearly impossible to remove the aphids between the florets through washing. Aphids tend to prefer the leaves of a broccoli plant to the stems; their presence there poses no problem since the leaves are not harvested. In high infestations, however, aphids spill over onto the flower head. Such produce will command a lower price or, in extreme infestation, will not even be harvested. Therefore, while broccoli yields per se may not be impacted by pest infestation, revenue most certainly is. The amount of “economic” damage can be inferred from the highest densities of aphids recorded in this study. Over the three-year study period, there were five instances of infestation of the head (aphid densities >1000/plant) in the week before harvest. All five of these instances occurred at sites with < 50% natural habitat; four of the five occurred at sites with < 12% natural habitat.
Regardless of how many of the above questions are answered in the affirmative, it is important to recognize that the study of landscape effects on pest control must look beyond its impact on natural enemies. Further, when considering natural enemy impact on pests, rate of pest population growth may provide a valuable metric in addition to pest densities. Several metrics should be included in any assessment of pest control by natural enemies, in order to determine the degree to which natural habitat can provide that ecosystem service and identify potential reasons why the service may be diminished. Improving our understanding of the function of natural enemies and the ecosystems that provide them will help us to manage our landscapes to maximize pest control services from nature, thus reducing or eliminating the use of harmful pesticides.
57
Chapter 5: Conclusion
Main findings
It is well established that habitat complexity enhances populations of beneficial insects on farms. The larger question, a more pressing question to growers and land managers, is the outcome of this enhancement on pest control and agricultural production. The goal of this dissertation research was to determine whether and under what conditions the natural enemies enhanced by habitat complexity can provide effective pest control on farms. The main findings can be addressed by revisiting the questions posed in the introduction:
Does habitat complexity provide pest control services to farms? The answer is a qualified yes. Habitat complexity, especially at the landscape scale, provides pest control services, but the apparent effect of those services on actual pest densities can be diminished by other mitigating factors. These factors may include the presence of host plants for pests discussed in Chapter Two, microclimate effects discussed in Chapters Three (temperature) and Four (precipitation), and factors not presented in this dissertation, such as dispersal patterns (Lewis 1965, Pasek 1988).
What aspects of habitat complexity provide pest control disservices to farms? Non-crop habitat can provide pests as well as enemies to crops. This research suggests that weeds growing around farm fields can serve not only as alternate host-plants but also as enemy-free space for crop pests. The goal of Chapter Two was to weigh the costs and benefits of mustard to aphids in order to assess its potential to serve as a source of crop pests, with the hypothesis that there would be some trade-off of aphid physiology for predation protection. The physiological experiments in this chapter indicate that by providing a predator-refuge without a substantial energetic cost to the pest, weedy mustard has greater potential than expected to serve as a source of aphids to surrounding broccoli crops. While the specifics of the mustard bomb may not be applicable to other (non-cole crop) systems, the lesson of natural habitat providing unexpected ecosystem disservices along with its services is an important one. If non-crop habitat around the farm has a high abundance of plants that are more beneficial to pests than to their enemies, it could potentially overwhelm any positive impact on pest control via natural enemies.
58 What is the role of different scales of habitat complexity in providing pest control services on farms? One goal of Chapter Three was to determine the importance of different scales of habitat complexity to pest control services, with the hypothesis that one scale could compensate for or enhance the other. Landscape complexity provides pest control services more consistently than local complexity; while local complexity appears capable of substituting for landscape complexity late in the season, early season trends suggest otherwise. A lag in enemy arrival weakens the overall impact of local complexity in simpler landscapes by potentially allowing pest populations to build up before their enemies arrive. Local complexity does not seem to augment landscape complexity, either, as no significant differences were detected between locally complex and locally simple farms in the more complex (natural-matrix) landscapes. A secondary goal of this chapter was to identify the conditions under which pest control services were maximized. The effects of cross- scale habitat complexity were examined with concurrent data on pest reproduction and colonization, which are influenced by temperature and the weedy refuges considered in Chapter Two. Warmer temperatures increase net aphid reproduction, which, along with presence of mustard, increase aphid colonization. Together, these data provide compelling evidence that habitat complexity at the landscape scale can enhance pest control services via top-down control, but suggest that bottom-up factors impacting pest populations must be considered as well.
How effective are pest control services at reducing pest densities on farms? While the cage experiment in Chapter Three documented enhanced pest control service related to habitat complexity, the sites receiving the highest levels of pest control service did not exhibit the lowest pest densities. The insect surveys in Chapter Four investigated why this might be the case, through multivariate analyses to test the hypothesis that landscape complexity improves syrphid abundance, aphid suppression, and in turn the provision of pest control services (as inferred from final aphid densities before harvest). The goal of this chapter was to determine to what degree natural habitat could deliver enemies to provide truly effective and reliable control. Landscape complexity increases syrphid abundance, which decreases aphid population growth. The expectation that pest population suppression by natural enemies from natural habitat should lead to a negative relationship between natural habitat and pest densities was validated to some extent. Pooling the data across all three years revealed a weak trend of pest abundance decreasing with natural habitat, and this relationship was made stronger by the inclusion of local complexity as an additional variable. It can be difficult to detect effects amidst so many other variables across time and space, which is presumably why the relationship between landscape and pests has proved so inconsistent in the literature. However, this dissertation suggests that the pest control services from natural habitat to farms are in fact significant in this system, but may not be detected unless temporal variation and the competing direct and indirect effects of landscape on pests are taken into account.
59 Future directions
Interactions between top-down and the additional bottom-up factors revealed here should be tested explicitly, by incorporating temperature and alternate host-plant distribution or phenology into models of the relationship between landscape complexity and pest densities or pest population change. These data could help improve our understanding of this relationship, much as accounting for inter-annual variation and local habitat differences did in our three-year insect survey.
Longer-term datasets would allow an investigation into broader concepts like resilience to pest infestation. Short-term responses like final aphid densities and suppression of aphid population growth may still not be the optimal metrics to capture the role of landscape complexity in pest dynamics. The beneficial insects provided by natural habitat in the landscape around farms may provide a form of “pest control insurance” to farms; pest densities in any given year may not be lower in more complex landscapes, but the frequency of periodic catastrophic pest outbreaks may be reduced. Finding proxies for pest infestation (such as insurance claims for losses filed with the federal government) and spatially referencing this information with respect to landscape complexity could be the next important step in discovering the true magnitude of services we receive from nature.
Implications of the research
The overall contribution of this dissertation is to shed light on why greater landscape complexity resulting in higher enemy abundance does not appear to reliably ensure lower pest populations on farms. Ecosystem services often go unnoticed until they are compromised or removed. Comparing one farm’s pest problems to that of another farm, with different management, microclimate, and location, is not a fair assessment of the contribution of the landscape around those farms to their pest control. Only through multi-year studies ranging over an extensive landscape gradient can we begin to tease apart the impact of habitat complexity from the many other variables influencing pest populations. The important message for growers and land-managers is that ecosystem services provided by natural habitat can play an important role in a broader pest management program, and habitat should be managed to maximize those services. Because agricultural landscapes currently occupy nearly 40% of the total land surface of the Earth, better management of this dominant habitat type would result in major benefits in improving food security, conserving biodiversity, and protecting human and ecosystem health.
60
References
Aliabadi A, Renwick JA, Whitman DW. 2002. Sequestration of glucosinolates by harlequin bug Murgantia histrionica. Journal of Chemical Ecology 28: 1749-62.
Andow D. 1990. Population dynamics of an insect herbivore in simple and diverse habitats. Ecology 71: 1006.
Basedow T. 1990. On the impact of boundary strips and of hedges on aphid predators, aphid attack and the necessity of insecticide applications in sugar beets. Gesunde Pflanzen 42: 241–245.
Beekwilder J, van Leeuwen W, van Dam NM, Bertossi M, Grandi V, Mizzi L, Soloviev M, Szabados L, Molthoff JW, Schipper B, Verbocht H, de Vos RC, Morandini P, Aarts MG, Bovy A. 2008. The impact of the absence of aliphatic glucosinolates on insect herbivory in Arabidopsis. PloS one 3: e2068.
Bianchi FJ, Booij CJ, 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-Biological Sciences 273: 1715-27.
Bianchi FJ, Goedhart PW, Baveco JM. 2008. Enhanced pest control in cabbage crops near forest in The Netherlands. Landscape Ecology 23: 595-602.
Bianchi F, van Wingerden W, Griffioen A, van der Veen M, van der Straten M, Wegman R, Meeuwsen H. 2005. Landscape factors affecting the control of by natural enemies in Brussels sprout. Agriculture, Ecosystems & Environment 107: 145- 150.
Boccaccio L, Petacchi R. 2009. Landscape effects on the complex of Bactrocera oleae parasitoids and implications for conservation biological control. BioControl 54: 607-616.
Bridges M, Jones AM, Bones AM, Hodgson C, Cole R, Bartlet E, Wallsgrove R, Karapapa VK, Watts N, Rossiter. JT. 2002. Spatial organization of the glucosinolate-myrosinase system in Brassica specialist aphids is similar to that of the host plant. Proceedings of the Royal Society B-Biological Sciences 269: 187- 191.
61 Cardinale BJ, Harvey CT, Gross K, Ives AR. 2003. Biodiversity and biocontrol: emergent impacts of a multi-enemy assemblage on pest suppression and crop yield in an agroecosystem. Ecology Letters 6: 857-865.
Chandler AE. 1968. The relationship between aphid infestations and oviposition by aphidophagous Syrphidae (Diptera). Annals of Applied Biology 61: 425-434.
Clough Y, Kruess A, Kleijn D, Tscharntke T. 2005. Spider diversity in cereal fields: comparing factors at local, landscape and regional scales. Journal of Biogeography 32: 2007-2014.
Cole RA. 1997. The relative importance of glucosinolates and amino acids to the development of two aphid pests Brevicoryne brassicae and Myzus persicae on wild and cultivated Brassica species. Entomologia Experimentalis et Applicata 85: 121-133.
Costamagna AC, Menalled FD, Landis DA. 2004. Host density influences parasitism of the armyworm in agricultural landscapes. Basic and Applied Ecology 5: 347-355.
Cui J, Jander G, Racki LR, Kim PD, Pierce NE, Ausubel FM. 2002. Signals involved in Arabidopsis resistance to Trichoplusia ni caterpillars induced by virulent and avirulent strains of the phytopathogen Pseudomonas syringae. Plant Physiology 129: 551-64.
DeBach P, Rosen D. 1991. Biological Control by Natural Enemies. London: Cambridge Univ. Press.
Den Belder E, Elderson J, Van Den Brink W, Schelling G. 2002. Effect of woodlots on thrips density in leek fields: a landscape analysis. Agriculture, Ecosystems & Environment 91: 139–145.
Denno R, Larsson S, Olmstead K. 1990. Role of enemy-free space and plant quality in host-plant selection by willow beetles. Ecology 71: 124–137.
Dixon A. 1977. Aphid ecology: life cycles, polymorphism, and population regulation. Annual Review of Ecology and Systematics 8: 329-353.
Dixon A. 1997. Aphid Ecology: An optimization approach. London: Chapman & Hall.
Dixon A, Howard M. 1986. Dispersal in aphids, a problem in resource allocation. In: Danthanarayana, W, Insect flight dispersal and migration. New York: Springer, p. 145–151.
Drapela T, Moser D, Zaller J, Frank T. 2008. Spider assemblages in winter oilseed rape affected by landscape and site factors. Ecography 31: 254-262.
62 Elliott N, Kieckhefer R, Beck D. 2002. Effect of aphids and the surrounding landscape on the abundance of Coccinellidae in cornfields. Biological Control 24: 214-220.
Elliott N, Kieckhefer R, Lee J, French B. 1999. Influence of within-field and landscape factors on aphid predator populations in wheat. Landscape Ecology 14: 239-252.
Elmqvist T, Folke C, Nyström M, Peterson G, Bengtsson J, Walker B, Norberg J. 2003. Response diversity, ecosystem change, and resilience. Frontiers in Ecology and the Environment 1: 488-494.
Fernandez-Cornejo J, Nehring R, Sinha E, Grube A, Vialou A. 2009. Assessing Recent Trends In Pesticide Use In U.S. Agriculture. Presented at the Annual Meeting of the Agricultural and Applied Economics Association (AAEA). 26–28 July 2009, AAEA, Milwaukee, WI. Available at Web site: http://purl.umn.edu/49271.
Fiedler AK, Landis DA, Wratten SD. 2008. Maximizing ecosystem services from conservation biological control: The role of habitat management. Biological Control 45: 254-271.
Francis F, Haubruge E, Gaspar C. 2000. Influence of host plants on specialist/generalist aphids and on the development of Adalia bipunctata (Coleoptera: Coccinellidae). European Journal of Entomology 97: 481-485.
Francis F, Lognay G, Wathelet JP, Haubruge E. 2001. Effects of allelochemicals from first (Brassicaceae) and second (Myzus persicae and Brevicoryne brassicae) trophic levels on Adalia bipunctata. Journal of Chemical Ecology 27: 243-256.
Francis F, Lognay G, Haubruge E. 2004. Olfactory responses to aphid and host plant volatile releases: (E)-beta-farnesene an effective kairomone for the predator Adalia bipunctata. Journal of Chemical Ecology 30: 741-755.
French BW, Elliott NC, Berberet RC, Burd JD. 2001. Effects of riparian and grassland habitats on ground beetle (Coleoptera: Carabidae) assemblages in adjacent wheat fields. Environmental Entomology 30: 225-234.
Gardiner MM, Landis DA, Gratton C, DiFonzo CD, O'Neal M, Chacon JM, Wayo MT, Schmidt NP, Mueller EE, Heimpel GE. 2009a. Landscape diversity enhances biological control of an introduced crop pest in the north-central USA. Ecological Applications 19: 143-54.
Gardiner MM, Landis DA, Gratton C, Schmidt N, O’Neal M, Mueller E, Chacon J, Heimpel GE, DiFonzo CD. 2009b. Landscape composition influences patterns of native and exotic lady beetle abundance. Diversity and Distributions 15: 554-564.
Gols R, Harvey JA. 2008. Plant-mediated effects in the Brassicaceae on the performance and behaviour of parasitoids. Phytochemistry Reviews 8: 187-206.
63
Gratton C, Welter SC. 1999. Does "enemy-free space" exist? Experimental host shifts of an herbivorous fly. Ecology 80: 773-785.
Gurr G, Wratten S. 2000. Biological Control: Measures of Success. Dordecht: Kluwer Academic Publishers.
Gurr GM, Wratten SD, Luna JM. 2003. Multi-function agricultural biodiversity: pest management and other benefits. Basic and Applied Ecology 4: 107-116.
Gurr GM, van Emden HF, Wratten S. 1998. Habitat manipulation and natural enemy efficiency: implications for control of pests. In: Barbosa, P, Conservation Biological Control. San Diego: Academic Press, 155-183.
Haenke S, Scheid B, Schaefer M, Tscharntke T, Thies C. 2009. Increasing syrphid fly diversity and density in sown flower strips within simple vs. complex landscapes. Journal of Applied Ecology 46: 1106-1114.
Hagler JR, Jackson. CG. 2001. Methods for marking insects: current techniques and future prospects. Annual Review of Entomology 46: 511-543.
Harvey JA, Gols R, Wagenaar R, Bezemer TM. 2007. Development of an insect herbivore and its pupal parasitoid reflect differences in direct plant defense. Journal of Chemical Ecology 33: 1556-69.
Hawkins BA, Cornell HV. 1994. Maximum parasitism rates and successful biological control. Science 266: 1886.
Heimpel G, Jervis M. 2005. Does floral nectar improve biological control by parasitoids? In: Wackers, F, van Rijn, P, Bruin, J, Plant-provided Food and Plant-carnivore Mutualism. Cambridge, UK.: Cambridge University Press, 267-304.
Holland J, Fahrig L. 2000. Effect of woody borders on insect density and diversity in crop fields: a landscape-scale analysis. Agriculture, Ecosystems \& Environment 78: 115–122.
Holzschuh A, Steffan-Dewenter I, Tscharntke T. 2008. Agricultural landscapes with organic crops support higher pollinator diversity. Oikos 117: 354-361.
Hopkins RJ, van Dam NM, van Loon JJ. 2009. Role of glucosinolates in insect-plant relationships and multitrophic interactions. Annual Review of Entomology 54: 57- 83.
Hughes RD. 1963. Population dynamics of the cabbage aphid Brevicoryne brassicae (L.). Journal of Animal Ecology 32: 393-424.
64 IPCS. 2002. Global assessment of the state-of-the-science of endocrine disruptors. World Health Organization, International Programme on Chemical Safety. Available at http://www.who.int/ipcs/publications/new_issues/endocrine_disruptors/en/
Isaia M, Bona F, Badino G. 2006. Influence of landscape diversity and agricultural practices on spider assemblage in Italian vineyards of Langa astigiana (Northwest Italy). Environmental Entomology 35: 297-307.
Jeffries MJ, Lawton JH. 1984. Enemy free space and the structure of ecological communities. Biological Journal of the Linnean Society 23: 269-286.
Jones VP, Hagler JR, Brunner JF, Baker CC, Wilburn TD. 2006. An inexpensive immunomarking technique for studying movement patterns of naturally occurring insect populations. Environmental Entomology 35: 827-836.
Jonsen I, Fahrig L. 1997. Response of generalist and specialist insect herbivores to landscape spatial structure. Landscape Ecology 12: 185–197.
Kazana E, Pope TW, Tibbles L, Bridges M, Pickett JA, Bones AM, Powell G, Rossiter JT. 2007. The cabbage aphid: a walking mustard oil bomb. Proceedings of the Royal Society B-Biological Sciences 274: 2271-7.
Kenmore PE. 1980. Ecology and outbreaks of a tropical insect pest of the green revolution, the rice brown planthopper, Nilaparvata lugens Stal. University of California, Berkeley.
Kim JH, Jander G. 2007. Myzus persicae (green peach aphid) feeding on Arabidopsis induces the formation of a deterrent indole glucosinolate. The Plant Journal 49: 1008-19.
Klein A, Steffan-Dewenter I, Tscharntke T. 2006. Rain forest promotes trophic interactions and diversity of trap-nesting Hymenoptera in adjacent agroforestry. The Journal of Animal Ecology 75: 315-23.
Kliebenstein D, Gershenzon J, Mitchell-Olds T. 2001. Comparative quantitative trait loci mapping of aliphatic, indolic and benzylic glucosinolate production in Arabidopsis thaliana leaves and seeds. Genetics 159: 359-370.
Kremen C, Chaplin-Kramer R. 2007. Insects as providers of ecosystem services: crop pollination and pest control. In: Stewart, AJ, New, TR, Lewis, OT, Insect Conservation Biology: Proceedings of the Royal Entomological Society's 23rd Symposium. Wallingford, UK: CABI Publishing, 349-382.
Kremen C, Williams NM, Aizen MA, Gemmill-Herren B, LeBuhn G, Minckley R, Packer L, Potts SG, Roulston T, Steffan-Dewenter I, Vázquez DP, Winfree R, Adams L, Crone EE, Greenleaf SS, Keitt TH, Klein A, Regetz J, Ricketts TH.
65 2007. Pollination and other ecosystem services produced by mobile organisms: a conceptual framework for the effects of land-use change. Ecology letters 10: 299- 314.
Kremen C, Williams NM, Bugg RL, Fay JP, Thorp RW. 2004. The area requirements of an ecosystem service: crop pollination by native bee communities in California. Ecology Letters 7: 1109-1119.
Kruess A. 2003. Effects of landscape structure and habitat type on a plant-herbivore- parasitoid community. Ecography 26: 283-290.
Kunert G, Otto S, Rose US, Gershenzon J, Weisser WW. 2005. Alarm pheromone mediates production of winged dispersal morphs in aphids. Ecology Letters 8: 596-603.
Landis DA, Wratten SD, Gurr GM. 2000. Habitat management to conserve natural enemies of arthropod pests in agriculture. Annual Review of Entomology 45: 175- 201.
Letourneau DK, Goldstein B. 2001. Pest damage and arthropod community structure in organic vs. conventional tomato production in California. Journal of Applied Ecology 38: 557 - 570.
Lewis T. 1965. The effects of an artificial windbreak on the aerial distribution of flying insects. Annals of Applied Biology 55: 503-512.
Losey JE, Denno RF. 1998. Positive predator-predator interactions: enhanced predation rates and synergistic suppression of aphid populations. Ecology 79: 2143.
Losey JE, Vaughan M. 2006. The economic value of ecological services provided by insects. BioScience 56: 311.
Menalled F, Costamagna A, Marino P, Landis D. 2003. Temporal variation in the response of parasitoids to agricultural landscape structure. Agriculture, Ecosystems & Environment 96: 29–35.
Menalled FD, Marino PC, Gage SH, Landis Da. 1999. Does Agricultural Landscape Structure Affect Parasitism and Parasitoid Diversity? Ecological Applications 9: 634-641.
Mithen R, Raybould AF, Giamoustairs A. 1995. Divergent selection for secondary metabolites between wild populations of Brassica oleracea and its implications for plant-herbivore interactions. Heredity 75: 472-484.
66 Muller C, Agerbirk N, Olsen C, Boeve J, Schaffner U, Brakefield P. 2001. Sequestration of host plant glucosinolates in the defensive hemolymph of the sawfly Athalia rosae. Journal of Chemical Ecology 27: 2505-2516.
Murphy SM. 2004. Enemy-free space maintains swallowtail butterfly host shift. Proceedings of the National Academy of Sciences of the United States of America 101: 18048-18052.
Naylor R, Ehrlich P. 1997. Natural pest control services and agriculture. In: Daily, G, Nature's Services. Washington: Island Press, 151-174.
Oberg S, Mayr S, Dauber J. 2008. Landscape effects on recolonisation patterns of spiders in arable fields. Agriculture, Ecosystems & Environment 123: 211-218.
Olmez-Bayhan S, Ulusoy M, Bayhan E. 2007. Is the parasitization rate of Diaeretiella rapae influenced when Brevicoryne brassicae feeds on Brassica plants? Phytoparasitica 35: 146-149.
Östman Ö, Ekbom B, Bengtsson J. 2001a. Landscape heterogeneity and farming practice influence biological control. Basic and Applied Ecology 2: 365–371.
Östman Ö, Ekbom B, Bengtsson J, Weibull AC. 2001b. Landscape complexity and farming practice influence the condition of polyphagous carabid beetles. Ecological Applications 11: 480-488.
PANNA. 2005. Cholinesterase Testing in California. Pesticide Action Network North America. Available at http://www.panna.org/resources/gpc/gpc_200504.15.1.05. dv.html
Pasek J. 1988. Influence of wind and windbreaks on local dispersal of insects. Agriculture, Ecosystems & Environment 22/23: 539-554.
Pimentel D, Acquay H, Biltonen M, Rice P, Silva M, Nelson J, Lipner V, Giordano S, Horowitz A, Damore M. 1992. Environmental and economic costs of pesticide use. Bioscience 42: 750-760.
Prasifka JR, Heinz KM, Minzenmayer RR. 2004. Relationships of landscape, prey and agronomic variables to the abundance of generalist predators in cotton (Gossypium hirsutum) fields. Landscape Ecology 19: 709-717.
Pratt C, Pope TW, Powell G, Rossiter JT. 2008. Accumulation of glucosinolates by the cabbage aphid Brevicoryne brassicae as a defense against two coccinellid species. Journal Of Chemical Ecology 34: 323-9.
67 Purtauf T, Roschewitz I, Dauber J, Thies C, Tscharntke T, Wolters V. 2005. Landscape context of organic and conventional farms: influences on carabid beetle diversity. Agriculture, Ecosystems & Environment 108: 165-174.
Ratzka A, Vogel H, Kliebenstein DJ, Mitchell-Olds T, Kroymann J. 2002. Disarming the mustard oil bomb. Proceedings of the National Academy of Sciences of the United States of America 99: 11223-8.
Rayor LS, Mooney LJ, Renwick JA. 2007. Predatory behavior of Polistes dominulus wasps in response to cardenolides and glucosinolates in Pieris napi caterpillars. Journal of Chemical Ecology 33: 1177-85.
Reichelt M, Brown P, Schneider B, Oldham N, Stauber E, Tokuhisa J, Kliebenstein D, Mitchell-Olds T, Gershenzon J. 2002. Benzoic acid glucosinolate esters and other glucosinolates from Arabidopsis thaliana. Phytochemistry 59: 663-671.
Ricketts TH, Daily GC, Ehrlich PR, Fay JP. 2001. Countryside biogeography of moths in a fragmented landscape: biodiversity in native and agricultural habitats. Conservation Biology 15: 378-388.
Rodman JE, Kruckeberg AR, Alshehbaz IA. 1981. Chemotaxonomic diversity and complexity in seed glucosinolates of Caulanthus and Streptanthus (Cruciferae). Systematic Botany 6: 197-222.
Root RB. 1973. Organization of a plant-arthropod in simple and diverse habitats: the fauna of collards (Brassica oleracea). Ecological Monographs 43: 95-124.
Roschewitz I, Hucker M, Tscharntke T, Thies C. 2005. The influence of landscape context and farming practices on parasitism of cereal aphids. Agriculture, Ecosystems and Environment 108: 218–227.
Rotheray GE. 1986. Colour, shape and defense in aphidophagous syrphid larvae (Diptera). Zoological Journal of the Linnean Society 88: 201-216.
Sarthou J, Ouin A, Arrignon F, Barreau G, Bouyjou B. 2005. Landscape parameters explain the distribution and abundance of Episyrphus balteatus (Diptera: Syrphidae). European Journal of Entomology: 539-545.
Schmidt MH, Tscharntke T. 2005. Landscape context of sheetweb spider (Araneae: Linyphiidae) abundance in cereal fields. Journal of Biogeography 32: 467-473.
Schmidt MH, Lauer A, Purtauf T, Thies C, Schaefer M, Tscharntke T. 2003. Relative importance of predators and parasitoids for cereal aphid control. Proceedings of the Royal Society B-Biological Sciences 270: 1905-1909.
68 Schmidt MH, Roschewitz I, Thies C, Tscharntke T. 2005. The differential effects of landscape and management on diversity and density of ground-dwelling farmland spiders. Journal of Applied Ecology 42: 281-287.
Schmidt M, Thies C, Nentwig W, Tscharntke T. 2008. Contrasting responses of arable spiders to the landscape matrix at different spatial scales. Journal of Biogeography 35: 157–166.
Schneider F. 1969. Bionomics and physiology of aphidophagous Syrphidae. Annual Review of Entomology 14: 103-124.
Settle WH, Ariawan H, Astuti ET, Cahyana W, Hakim AL, Hindayana D, Lestari AS, Pajarningsih, Sartanto. 1996. Managing tropical rice pests through conservation of generalist natural enemies and alternative prey. Ecology 77: 1975-1988.
Snyder WE, Ives AR. 2003. Interactions between specialist and generalist natural enemies: parasitoids, predators, and pea aphid biocontrol. Ecology 84: 91-107.
Stearns S. 1992. The evolution of life histories. Oxford: Oxford University Press.
Steffan-Dewenter I. 2002. Landscape context affects trap-nesting bees, wasps, and their natural enemies. Ecological Entomology 27: 631–637.
Thackray DJ, Diggle AJ, Berlandier FA, Jones RA. 2004. Forecasting aphid outbreaks and epidemics of Cucumber mosaic virus in lupin crops in a Mediterranean-type environment. Virus research 100: 67-82.
Thies C, Tscharntke T. 1999. Landscape structure and biological control in agroecosystems. Science 285: 893-895.
Thies C, Roschewitz I, Tscharntke T. 2005. The landscape context of cereal aphid- parasitoid interactions. Proceedings of the Royal Society B-Biological Sciences 272: 203-10.
Thies C, Steffan-Dewenter I, Tscharntke T. 2003. Effects of landscape context on herbivory and parasitism at different spatial scales. Oikos 101: 18–25.
Thies C, Steffan-Dewenter I, Tscharntke T. 2008. Interannual landscape changes influence plant–herbivore–parasitoid interactions. Agriculture, Ecosystems & Environment 125: 266-268.
Tscharntke T, Bommarco R, Clough Y, Crist TO, Kleijn D, Rand TA, Tylianakis JM, Nouhuys SV, Vidal S. 2007. Conservation biological control and enemy diversity on a landscape scale. Biological Control 43: 294-309.
69 Tscharntke T, Klein AM, Kruess A, Steffan-Dewenter I, Thies C. 2005. Landscape perspectives on agricultural intensification and biodiversity – ecosystem service management. Ecology Letters 8: 857-874. Tscharntke T, Steffan-Dewenter I, Kruess A, Thies C. 2002. Contribution of small habitat fragments to conservation of insect communities of grassland–cropland landscapes. Ecological Applications 12: 354-363.
Ulusoy M, Olmez-Bayhan S. 2006. Effect of certain Brassica plants on biology of the cabbage aphid Brevicoryne brassicae under laboratory conditions. Phytoparasitica 34: 133-138.
Valantin-Morison M, Meynard J, Dore T. 2007. Effects of crop management and surrounding field environment on insect incidence in organic winter oilseed rape (Brassica napus L.). Crop Protection 26: 1108-1120. van Emden H. 1963. A field technique for comparing the intensity of mortality factors acting on the Cabbage aphid, Brevicoryne brassicae (L.) (Hem: Aphididae), in different areas of a crop. Entomologia Experimentalis et Applicata 6: 53–62.
Vanhaelen N, Gaspar C, Francis F. 2002. Influence of prey host on a generalist aphidophagous predator: Episyrphus balteatus (Diptera: Syrphidae). European Journal of Entomology 99: 561-564.
Vollhardt IM, Tscharntke T, Wäckers FL, Bianchi FJ, Thies C. 2008. Diversity of cereal aphid parasitoids in simple and complex landscapes. Agriculture, Ecosystems & Environment 126: 289-292.
Werling BP, Gratton C. 2008. Influence of field margins and landscape context on ground beetle diversity in Wisconsin (USA) potato fields. Agriculture, Ecosystems & Environment 128: 104-108.
Winkler K, Wäckers F, Pinto DM. 2009. Nectar-providing plants enhance the energetic state of herbivores as well as their parasitoids under field conditions. Ecological Entomology 34: 221-227.
Zaller JG, Moser D, Drapela T, Schmöger C, Frank T. 2008. Insect pests in winter oilseed rape affected by field and landscape characteristics. Basic and Applied Ecology 9: 682-690.
Zaller JG, Moser D, Drapela T, Schmöger C, Frank T. 2009. Parasitism of stem weevils and pollen beetles in winter oilseed rape is differentially affected by crop management and landscape characteristics. BioControl 54: 505-514.
Zehnder G, Gurr GM, Kühne S, Wade MR, Wratten SD, Wyss E. 2007. Arthropod pest management in organic crops. Annual Review of Entomology 52: 57-80.
70