Urban agriculture occurs within and near the built environment, with high proportions of surrounding impervious surfaces such as buildings and roads . Urbanization fragments habitats and reduces the abundance and diversity of organisms . Many of the affected organisms are beneficial to urban agriculture and the provision of ecosystem services such as pollination and biological control services . Biological control of pest insects is an important ecosystem service for urban farmers due to pesticide use regulation in cities and rejection of chemical management practices for health and environmental reasons. In the absence of chemical controls, agroecological pest management practices are frequently adopted. Agroecological pest management is a proactive ecosystem services based approach that aims to reduce pest abundance and crop damage by increasing natural enemy populations through agroecosystem diversification . However, landscape fragmentation and surrounding imperviousness can often negatively affect the regulatory ecosystem services APM relies upon . While most off-farm landscape effects are not within the control of urban farmers, on-farm diversification practices are. Substantial research and published literature have investigated the impact of diversification practices to increase biological control of pest insects on rural farms, but less attention has been focused on the effects of diversification in the fragmented landscapes innate to urban agroecosystems . This chapter focuses on the effects of floral provisioning on parasitic Hymenoptera and the ubiquitous cabbage aphid , specifically,30 plant pot the impact of floral provisioning on PH populations vis-a-vis the enemies hypothesis and the nectar provision hypothesis .
The nectar provision hypothesis was proposed to explore the effects of floral-based diversification schemes on the contributions made by PH to biological control services. Heimpel and Jervis posit that with increased accessibility to nectar-producing plants, PH should respond with increased fitness, resulting in elevated levels of localized biological control . Many PH species have been documented feeding on a wide variety of flowers . Nectar, pollen, and extrafloral nectar are essential sources of carbohydrates, proteins, lipids, and minerals for PH . Nectar provisioning has been shown to play an important role in increasing parasitoid longevity, fecundity , abundance, and diversity , as well as increase rates of parasitism . Despite documented positive outcomes for parasitoid fitness and increased rates of biological control, floral provisioning does not always result in improved biological control services from PH . In response to confounding results regarding floral manipulations, researchers have proposed several concepts to explain these inconsistencies: 1. PH may utilize floral resources but then disperse to reduce the risk of hyper- or super-parasitism, other mortality, and inbreeding among offspring ; 2. parasitoids may already have enough local floral resources, and floral manipulations may not be introducing a limited resource ; 3. pest insects utilize floral resources more effectively than parasitoids ; 4. diversification strategies might make it difficult for parasitoids to find hosts in increasingly heterogeneous landscapes ; and lastly, 5. many factors determine the ability of PH to use floral resources, including wasp body size, mouthpart morphology, floral structure, and nutritional value . A disconnect between plant species and parasitoid feeding characteristics may limit the opportunity of PH to utilize these floral resources . The extent to which these conditions affect PH in urban areas is still being explored.
Generally speaking, the inclusion of flowers into urban agroecosystems to supply nectar for PH should yield effective results in the context of APM. However, inconsistent results regarding the nectar provisioning hypothesis and effects on biological control services have complicated the implementation of floral provisioning practices for farmers. Of all potential remedies for inconsistency regarding effects of floral provisioning, the concept of functional biodiversity has been championed for its potential to influence habitat manipulations that are targeted toward specific ecosystem services or to particular natural enemies. Understanding the linkages between potential PH feeding preferences and specific agroecosystem components could help farmers “fine-tune” their production systems to maximize biological control services. Morphology, bloom time, floral area, and the amount of pollen and nectar resources provided by a given plant species have all been shown to either positively or negatively impact natural enemy populations . Gaining a better understanding of the range of flowers most likely to be utilized by and positively affect PH populations and biological control services may enable practitioners to tailor management practices . To better understand the effects of floral provisioning on PH richness and abundance and potential feeding preferences, we conducted an in-situ flower survey using an improvised D-vac insect vacuum fitted with a lined and filtered five-gallon bucket that wholly covered flowering plant inflorescences. Each sampled plant was visually assessed for spatial relationships regarding other herbaceous cover and was only sampled if it was standing free of additional herbaceous cover and flowering plants. In addition, each plant was visually assessed for pest infestations and was not sampled if pest infestations were visible. We vacuumed three plants of each flowering plant species present at a farm location. Multiple samples were collected during a farm visits, but varied due to available specimens.
Sampling occurred once every thirty days during the same time intervals during each visit. Results from the 2018 survey informed flowering plant selection for the 2019 sampling season. Each plant species that yielded very few or no PH during the 2018 sampling period was excluded from sampling in the following year; 13 flowering plant species were sampled . Each sample was stored in a deep freeze until processed by extracting all PH and identifying them to sub-family as per previous literature . Parasitic Hymenoptera identification was accomplished using Goulet et al., for all groups and Gibson et al. for Chalcidoidea and Dangerfield et al., for Braconidae. Collected specimens that were damaged were identified to morphosubfamily.Aphid abundance, parasitism, and plant damage observations were performed over two growing seasons on commonly grown brassica cultivars: kale, broccoli, collards, and tree collards. Individual plants were randomly selected and identified to cultivar. If possible, we only observed plants that had not been heavily harvested. The major insect pests of interest were cabbage aphids , a common agricultural pest of Brassicaceae. Aphid abundance was measured on each plant by selecting three leaves and recording the number of apterous, alate, and parasitized aphids . The percent of mummified aphids per leaf was used as a measure of biological control services by parasitoid wasps. A qualitative assessment of pest damage on brassicas was completed using a high, medium, and low scale based on familiar concepts of marketability. High damage corresponded to a leaf that would be unmarketable, medium damage had some damage but would still be purchased by a consumer, and low damage had little to no visible damage. Two agroecological practices that increase on-farm diversification, floral provisioning,30 planter pot and crop richness were measured three times: early-season , mid-season , and late-season . Crop richness was measured by using 8m transects across cropping systems. Any crop plant that touched the transect line was considered, including different cultivars of the same species . Three transects were completed on small farms, six on medium farms, and nine on large farms. We collected data for crop richness three times during the growing season over the two years of the study. Floral richness was recorded seasonally, similar to crop richness. Every on-farm, non-crop flowering plant was recorded and identified to genus. Generalized linear mixed models were constructed for each of the following response variables: Total parasitic Hymenopteran abundance, super-family, family, and subfamily abundance, and total PH diversity. Selected fixed effect explanatory variables included: floral richness, floral species, year by season, and site as a random intercept. Using the fitdistrplus package in R, parasitic Hymenoptera count data were plotted and examined to determine the best probabilistic distribution for the GLMM modeling; a Poisson distribution with a log link function . Models that had response variables significantly affected by floral species were further analyzed using the non-parametric KruskalWallis test with posthoc Dunn’s test to determine floral species that had the most significant influence on the response variable . Aphid data were analyzed to test for differences in aphid abundance, parasitism rates, and crop damage.
Explanatory variables examined were year and season, floral richness?Aphid count data were assessed using the fitdistrplus package in R to determine the best probabilistic distribution for the GLMM modeling; a negative binomial distribution with a log link function . The final GLMM was constructed with glmer.nb using crop and floral richness, date and year as fixed effects and site as the random effect. We constructed mixed-effects models using the lme4 package in R . After fitting a series of GLMMs based on predictors expected to affect response variables, the model with the lowest Akaike Information Criterion score was selected. All GLMM model residuals were simulated from the fitted model using the simulate Residuals function in the package DHARMa to test for dispersion and model fit . Using the effects package in R, a partial regression plot was constructed for each predictor variable included in the final GLMMs . To better understand the nectar provision hypothesis and the effectiveness of floral-based habitat manipulations, we used PH abundance and richness on flowering plants to indicate parasitoid feeding preferences. Our results showed that floral species were not a strong indicator of increased abundance or richness at any measured scale of PH. Only one PH family, the pteromalids, was found in more significant quantities on nettles. Using a non-parametric test, PH diversity was shown to increase on marigolds and nettles, a response documented by previous floral provisioning research . Laboratory experiments have shown floral feeding preferences in parasitoids , but in-situ results have been less clear . Our results show that our collected PH had a very weak response to floral species, and in some cases showed a negative relationship to floral richness. Several factors may singularly, or in aggregate, explain this absence of floral preference in situ: 1. floral resources incorporated into urban gardens and farms are not selected for their functional diversity but that of other traits, such as attractiveness and availability ; 2. in-situ food resources may present a greater variety of acceptable foods unlike no-choice feeding trials; and 3. some sampling bias may have occurred when using the vacuum on inflorescences as the vacuum may be more likely to capture smaller parasitoids which may be able to exploit a broader range of nectaries or may be feeding on other food items such as honeydew . Many parasitoids also feed on the same plant as their host, which may bias visitation by family and sub-family. A negative PH abundance response to increased floral richness may be a result of dispersion to reduce the risk of hyper- or super-parasitism, other mortality, and inbreeding among offspring . Our research showed a weak relationship between increased PH richness on marigolds and nettles and an increased abundance response with pteromalids on nettles compared to other PH taxa. Marigolds have a history of being utilized as a beneficial flower in the gardening community and are grown for cultural and aesthetic reasons. Nettles are not typically grown intentionally, and in the few location’s nettles were sampled, they were unintentional but preserved in non-crop areas. Nettles, in this case, may be an example of a non-selected floral species with a higher level of ecosystem function than species selected for other traits. Additionally, aphid parasitoids, specifically Aphidius, have been found in higher abundance on nettles due to the occurrence of the stinging nettle aphid, Microlophiurn carnosum . It is unclear what connection pteromalids may have in this ecology. It is possible that nettles sampled had infestations of aphids that remained obscured due to the obstacles associated with close inspection of stinging nettles. Despite these findings, anecdotal relationships between floral species and specific PH species indicate these relationships should continue to be explored to better understand parasitoid feeding preferences, floral occupancy, and farm scape mediated biological control. To assess the second criterion of the nectar provision hypothesis, a demonstratable reduction in pest impacts, we looked at aphid abundance, rates of parasitism, and overall crop damage on brassicas. Our results show that farms with increased floral richness have lower aphid counts per plant. We did not record a reduction in crop damage nor an increase in aphid parasitism with increased floral richness.