Evaluating Pollinator-mediated Seed Set in Urban Native Flora Communities

Thesis Abstract

Urban ecosystems are increasingly characterized by habitat fragmentation, invasive species pressure, and shifting pollinator communities, which together threaten the reproductive success of native flora. This study addresses the problem of uncertain seed set outcomes for urban native plant assemblages amid anthropogenic pressures, aiming to quantify pollinator-mediated seed production and identify drivers of reproductive success in city landscapes. The objectives are to (i) quantify seed set and fruiting success in a representative array of urban native species across contrasting habitat patches, (ii) evaluate pollinator visitation rates, foraging behavior, and pollinator diversity as predictors of seed set, (iii) compare seed set outcomes among parkland, street verge, and remnant habitats, and (iv) examine the relative influence of floral traits, plant phenology, and neighborhood green infrastructure on pollination effectiveness. A mixed-methods approach was employed, integrating quantitative field experiments with observational pollinator surveys conducted across three metropolitan areas during two flowering seasons (spring and late summer). The study population comprises 24 native herbaceous and shrub species common to urban corridors, sampled from 60 plots stratified by habitat type and isolation index. Within each plot, controlled pollination treatments (open pollination, manual self-pollination, and emasculated pollination) were applied to estimate natural pollination reliance and self-compatibility. Data collection instruments included standardized seed counts per capsule, fruit set proportions, and seed viability assays, complemented by timed transects to record pollinator visitation (Apis mellifera, Bombus spp., Andrena spp., and other native taxa) and floral trait measurements (petal color, corolla depth, nectar volume, and bloom duration). Environmental covariates such as soil moisture, urban heat island intensity, and local vegetation complexity were captured using portable sensors and high-resolution GIS derived metrics. Analytical procedures integrated generalized linear models (GLMs) and generalized additive models (GAMs) to relate seed set to pollinator metrics, habitat type, and floral traits, while controlling for plant species and year effects. Multilevel mixed-effects models accounted for plot-level nesting, and Poisson or binomial error structures were applied as appropriate for seed counts and proportion data. Structural equation modeling (SEM) tested causal pathways among habitat fragmentation, pollinator diversity, visitation rate, and seed set. Diversity analyses (Shannon–Wiener index) quantified pollinator assemblages, and a partial least squares regression (PLSR) identified the most influential predictors of seed set across species. Sensitivity analyses assessed the robustness of results to sampling intensity and spatial autocorrelation. Ethical approvals and permits were secured for field manipulation and pollinator observations, with data management following open science principles. Expected findings anticipate higher seed set in plots with diverse pollinator communities and frequent visitation, particularly in patches with moderate connectivity and abundant nectar resources. It is hypothesized that open pollination will yield seed sets closer to natural conditions, while emasculated controls will reveal the degree of self-compatibility across species. Species with longer bloom duration and richer nectar rewards are expected to demonstrate higher pollination success, yet urban stressors such as heat and soil compaction may dampen these effects in highly fragmented habitats. The study is likely to reveal significant interactions between habitat type and pollinator diversity, with remnant and parkland patches supporting more robust pollination services than street verge habitats. The contribution to knowledge lies in providing empirical, site-specific estimates of how urban pollinator dynamics translate into seed production for native flora, informing urban biodiversity planning, green infrastructure design, and conservation priorities. The research advances theory by integrating pollination ecology with urban landscape ecology, offering a scalable framework to evaluate pollination effectiveness in heterogeneous city environments. Practical implications include guidance for selecting locally adapted native species with favorable pollination syndromes, optimizing floral resource mosaics across urban matrices, and prioritizing restoration in patches that maximize pollinator access. The study concludes that enhancing habitat connectivity and augmenting floral resource diversity within urban green networks are critical to sustaining seed set and, by extension, native plant persistence under ongoing urbanization. Recommendations emphasize targeted management of pollinator habitats, routine monitoring of pollinator assemblages, and incorporation of pollination metrics into urban planning and biodiversity certification programs.