Study Notes: Wild Bee Pesticide Exposure in Urban Grasslands and Community Gardens

Pesticide Exposure in Urban Wild Bee Genera

• Global wild bee declines pose severe risks to sustainable food production and natural ecosystem functioning.

• Conventional agricultural landscapes rely heavily on chemical inputs (insecticides, fungicides, and herbicides) and offer reduced floral diversity, driving pollinator loss.

• Urban environments frequently contain high floral abundance across green spaces (residential gardens, community gardens, wildlife refuges, and parks), supporting high pollinator species richness relative to surrounding agricultural land.

• Potential urban pesticide exposure pathways for wild pollinators include:

  • Residential lawn and garden treatments ("bug sprays", fungicides, and weed killers) applied by untrained private users.

  • Pesticide drift onto surrounding wild flowering plants from treated private lawns or gardens.

  • Ornamental plants purchased from commercial retail nurseries containing systemic agrochemical residues in nectar and pollen.

  • Veterinary antiparasitic treatments (such as topical flea treatments containing fipronil or imidacloprid) entering urban watersheds or floral surfaces.

  • Public health vector management programs (such as aerosol spraying targeting mosquitoes).

• Across $8$ bee genera foraging in urban landscapes, $13$ unique pesticide compounds were detected, including insecticides, fungicides, herbicides, and a synergist, with exposure frequencies and residue concentrations strongly correlated with bee body size and foraging range.

Field Collection and Sampling Methodology

• Field sampling was conducted between May 22, 2020, and June 9, 2020, across $10$ urban sites located in Travis County, Texas, USA.

• The $10$ sites comprised $5$ community gardens and $5$ urban grassland patches:

  • Community gardens ranged in area from $1885.82\,m^2$ to $17710.10\,m^2$. Pesticide application was strictly prohibited in all sampled community gardens.

  • Grassland sites ranged in area from $3867.13\,m^2$ to $740893.67\,m^2$. Grassland sites were managed primarily as restored prairie habitat without direct pesticide applications.

  • All sites were separated by a spatial distance of at least $1\,km$ to ensure spatial independence among foraging bee populations.

• Sampling was conducted during morning hours ($7:00\text{--}12:00$) to coincide with peak daily pollinator activity.

• Actively foraging bees were captured directly off flowers using aerial nets, transferred immediately to individual collection vials, and stored on ice packs in the field before long-term storage at $-80\,^\circ\text{C}$.

• A total of $322$ individual bees were collected across $8$ target genera representing diverse life histories, nesting habits, and social structures:

  • Agapostemon ($n = 34$ individual bees)

  • Apis mellifera ($n = 46$ individual bees)

  • Halictus ($n = 39$ individual bees)

  • Lasioglossum ($n = 55$ individual bees)

  • Megachile ($n = 32$ individual bees)

  • Melissodes ($n = 50$ individual bees)

  • Osmia ($n = 28$ individual bees)

  • Xylocopa ($n = 38$ individual bees)

• To provide adequate biological mass for analytical chemistry screening, smaller and less abundant bee species were pooled by genus at each site, yielding $181$ analyzed analytical samples:

  • Agapostemon: $n = 10$ pooled samples

  • Halictus: $n = 9$ pooled samples

  • Lasioglossum: $n = 10$ pooled samples

  • Megachile: $n = 10$ pooled samples

  • Osmia: $n = 8$ pooled samples

  • Apis mellifera: $n = 46$ individual samples

  • Melissodes: $n = 50$ individual samples

  • Xylocopa: $n = 38$ individual samples

Analytical Chemistry and Extraction Protocol

• Chemical analysis screened for $92$ distinct pesticide active ingredients, breakdown products, and metabolites using Liquid Chromatography-Mass Spectrometry (LC-MS/MS).

• Reagents and Solvents:

  • HPLC-grade acetonitrile and water.

  • LC-MS grade formic acid.

  • $5\,M$ ammonium formate solution.

  • Agilent QuEChERS extraction packets containing $4\,g$ $\text{MgSO}_4$, $1\,g$ $\text{NaCl}$, $1\,g$ sodium citrate tribasic dihydrate, and $0.5\,g$ sodium citrate dibasic sesquihydrate.

  • Dispersive solid-phase extraction (d-SPE) kits containing $150\,mg$ $\text{MgSO}_4$, $25\,mg$ primary-secondary amine (PSA), and $25\,mg$ C18EC.

  • Deuterated internal standards: $d_4\text{-fluopyram}$, $d_3\text{-pyraclostrobin}$, and $^{13}C_6\text{-metalaxyl}$.

• Sample Preparation Protocol (Modified EN 15662 QuEChERS Method):

  • Whole frozen bee samples weighing between $0.079\,g$ and $0.561\,g$ were hydrated with $300\,\mu L$ of water and combined with $700\,\mu L$ of acetonitrile.

  • Homogenization was executed for $1\,min$ using $2.8\,mm$ diameter ceramic beads on a Bead Ruptor 24 bead mill.

  • Large individual specimens (Xylocopa and Megachile) exceeding $600\,mg$ ($n = 8$) were cryogenically ground in liquid nitrogen using a mortar and pestle; $500\,mg\,(\pm 1\%)$ of pulverized tissue was combined with $300\,\mu L$ water and $700\,\mu L$ acetonitrile, then vortexed for $1\,min$.

  • Following homogenization, $330\,mg$ of EN 15662 salts ($203\,mg$ $\text{MgSO}_4$, $51\,mg$ $\text{NaCl}$, $51\,mg$ sodium citrate tribasic dihydrate, $25\,mg$ sodium citrate dibasic sesquihydrate) were added to induce phase separation.

  • Samples were vortexed thoroughly and centrifuged at $7300 \times g$ for $5\,min$.

  • A $1\,mL$ volume of the upper organic supernatant was transferred to a d-SPE tube ($150\,mg$ $\text{MgSO}_4$, $25\,mg$ PSA, $25\,mg$ C18EC) for matrix clean-up.

  • From the cleaned extract, $297\,\mu L$ of supernatant was harvested and mixed with $3\,\mu L$ of internal standard solution ($0.15\,\mu g/mL$ $d_4\text{-fluopyram}$, $0.3\,\mu g/mL$ $d_3\text{-pyraclostrobin}$, $0.3\,\mu g/mL$ $^{13}C_6\text{-metalaxyl}$).

  • Extracts were filtered through $0.22\,\mu m$ PTFE membrane filters prior to injection.

• Liquid Chromatography and Mass Spectrometry Instrumentation Parameters:

  • UHPLC System: Dionex Softron Vanquish Flex UHPLC coupled to a Thermo Scientific TSQ Quantis mass spectrometer.

  • Chromatographic Column: Waters Acquity UPLC BEH C18 column ($100\,mm \times 2.1\,mm$, $1.7\,\mu m$ particle size) maintained at a constant temperature of $40\,^\circ\text{C}$.

  • Flow Rate: $300\,\mu L/min$.

  • Mobile Phase Components:

  • Mobile Phase A: Water containing $2\,mM$ ammonium formate and $0.1\%$ formic acid.

  • Mobile Phase B: Acetonitrile/Water ($98:2$, v/v) containing $2\,mM$ ammonium formate and $0.1\%$ formic acid.

  • Gradient Elution Schedule:

  • $1.5\,min$ initial equilibration ($2\%$ B)

  • $0\text{--}0.5\,min$: $2\%$ B (isocratic)

  • $0.5\text{--}15\,min$: $2\% \rightarrow 70\%$ B (linear gradient)

  • $15\text{--}17\,min$: $70\% \rightarrow 100\%$ B (linear gradient)

  • $17\text{--}20\,min$: $100\%$ B (column wash)

  • $20\text{--}20.2\,min$: $100\% \rightarrow 2\%$ B (linear gradient)

  • $20.2\text{--}23\,min$: $2\%$ B (re-equilibration)

  • Heated Electrospray Ionization (H-ESI) Parameters:

  • Spray voltage: $2000\,V$ (both positive and negative ion modes)

  • Sheath gas: $55$ arbitrary units

  • Auxiliary gas: $25$ arbitrary units

  • Sweep gas: $2$ arbitrary units

  • Ion transfer tube temperature: $325\,^\circ\text{C}$

  • Vaporizer temperature: $350\,^\circ\text{C}$

  • Mass Spectrometric Mode: Selected Reaction Monitoring (SRM) with two transitions per targeted analyte (one quantification transition, one confirmation transition).

  • Q1 and Q3 resolution set at $0.7$ Full Width at Half Maximum (FWHM); cycle time set at $0.4\,s$; argon collision gas pressure set at $2\,mTorr$.

• Data Classification and Statistical Modeling:

  • Chemical residue levels were categorized into three tiers:

  • Undetected compounds: residue value set to $0\,ppb$.

  • Trace compounds (detected below quantifiable limits): estimated concentration assigned as the compound-specific lowest limit of detection ($\text{LOD}$).

  • Quantifiable compounds: assigned the absolute quantified value in $ppb$

  • Generalized linear models (GLMs) with a binomial error structure were employed to evaluate pesticide detection probability across samples. Full models evaluated genus, sampling site, and mean sample bee mass. Model selection based on Akaike Information Criterion adjusted for small sample sizes ($\Delta\text{AICc} < 2$) retained bee genus as the sole significant explanatory factor ($p < 0.001$).

Pesticide Detection Frequencies and Residue Levels

• Overall Pesticide Occurrence:

  • $13$ out of $92$ screened pesticide active ingredients were detected in wild and managed urban bees.

  • The detected spectrum contained $5$ insecticides, $4$ fungicides, $3$ herbicides, and $1$ insecticide synergist.

  • Across all $181$ analyzed analytical samples, detection frequencies were:

  • Piperonyl butoxide (synergist): $8.3\%$

  • Atrazine (herbicide): $3.3\%$

  • Fipronil (insecticide): $2.8\%$

  • Diuron (herbicide): $2.2\%$

  • Thiamethoxam (insecticide): $2.2\%$

  • Carbaryl (insecticide): $1.1\%$

  • Dinotefuran (insecticide): $0.6\%$

  • Imidacloprid (insecticide): $0.6\%$

  • 4-Hydroxy-chlorothalonil (fungicide metabolite): $0.6\%$

  • Azoxystrobin (fungicide): $0.6\%$

  • Tebuconazole (fungicide): $0.6\%$

  • Metolachlor (herbicide): $0.6\%$

  • Difenoconazole (fungicide): $0.6\%$

• Genus-Specific Exposure Patterns:

  • Xylocopa: $52\%$ of samples contained at least one chemical; $11$ total distinct pesticide compounds were detected in this genus.

  • Megachile: $50\%$ of samples contained detected pesticides; $3$ distinct compounds were detected.

  • Osmia: $12.5\%$ of samples contained detected pesticides; $1$ compound detected.

  • Apis mellifera: $10.8\%$ of samples contained detected pesticides; $4$ distinct compounds detected.

  • Melissodes: $10.0\%$ of samples contained detected pesticides; $4$ distinct compounds detected.

  • Agapostemon, Halictus, and Lasioglossum: $0.0\%$ detection frequency across all samples ($0$ pesticide compounds detected).

  • Co-occurrence of multiple pesticides in a single sample occurred exclusively in Xylocopa ($6$ total samples; $5$ samples contained $2$ pesticides, $1$ sample contained $3$ pesticides).

• Compound Quantification Summary Across Bee Genera:

  • Carbaryl:

  • Apis mellifera: Mean = $0.01\,ppb$, Max = $0.52\,ppb$

  • Xylocopa: Mean = $0.01\,ppb$, Max = $0.34\,ppb$

  • All other genera: Undetected ($\text{ND}$)

  • Dinotefuran:

  • Melissodes: Mean = $0.05\,ppb$, Max = $2.74\,ppb$

  • All other genera: Undetected ($\text{ND}$)

  • Fipronil:

  • Megachile: Mean = $0.04\,ppb$, Max = $0.36\,ppb$

  • Xylocopa: Mean = $0.81\,ppb$, Max = $18.51\,ppb$

  • All other genera: Undetected ($\text{ND}$)

  • Imidacloprid:

  • Xylocopa: Mean = $0.02\,ppb$, Max = $0.81\,ppb$

  • All other genera: Undetected ($\text{ND}$)

  • Thiamethoxam:

  • Apis mellifera: Mean = $0.02\,ppb$, Max = $0.51\,ppb$

  • Melissodes: Mean = $0.03\,ppb$, Max = $1.31\,ppb$

  • Xylocopa: Mean = $0.02\,ppb$, Max = $0.88\,ppb$

  • All other genera: Undetected ($\text{ND}$)

  • Azoxystrobin:

  • Xylocopa: Mean = $0.00\,ppb$, Max = $0.04\,ppb$

  • All other genera: Undetected ($\text{ND}$)

  • Difenoconazole:

  • Xylocopa: Mean = $0.01\,ppb$, Max = $0.52\,ppb$

  • All other genera: Undetected ($\text{ND}$)

  • Tebuconazole:

  • Apis mellifera: Mean = $0.04\,ppb$, Max = $1.94\,ppb$

  • All other genera: Undetected ($\text{ND}$)

  • 4-Hydroxy-chlorothalonil:

  • Xylocopa: Mean = $0.59\,ppb$, Max = $22.25\,ppb$

  • All other genera: Undetected ($\text{ND}$)

  • Atrazine:

  • Apis mellifera: Mean = $0.00\,ppb$, Max = $0.20\,ppb$

  • Megachile: Mean = $0.00\,ppb$, Max = $0.03\,ppb$

  • Xylocopa: Mean = $0.02\,ppb$, Max = $0.38\,ppb$

  • All other genera: Undetected ($\text{ND}$)

  • Diuron:

  • Melissodes: Mean = $0.02\,ppb$, Max = $1.21\,ppb$

  • Xylocopa: Mean = $0.07\,ppb$, Max = $1.90\,ppb$

  • All other genera: Undetected ($\text{ND}$)

  • Metolachlor:

  • Xylocopa: Mean = $0.04\,ppb$, Max = $1.45\,ppb$

  • All other genera: Undetected ($\text{ND}$)

  • Piperonyl Butoxide:

  • Megachile: Mean = $0.01\,ppb$, Max = $0.05\,ppb$

  • Melissodes: Mean = $0.02\,ppb$, Max = $1.08\,ppb$

  • Osmia: Mean = $0.01\,ppb$, Max = $0.06\,ppb$

  • Xylocopa: Mean = $0.05\,ppb$, Max = $0.73\,ppb$

  • All other genera: Undetected ($\text{ND}$)

Drivers and Implications of Urban Agrochemical Exposure

• Influence of Body Size and Foraging Radius:

  • Average wet body masses per genus were:

  • Xylocopa: $\sim 355\,mg$

  • Megachile: $\sim 105\,mg$

  • Melissodes: $\sim 40\text{--}60\,mg$

  • Agapostemon: $\sim 33\,mg$

  • Halictus: $\sim 20\,mg$

  • Lasioglossum: $\sim 4\,mg$

  • Maximum foraging distances scale non-linearly with bee body size. Smaller bees (Lasioglossum, Halictus, Agapostemon) possess restricted foraging radii, confining their foraging activity to the immediate boundaries of the sampled pesticide-free community gardens and grassland restorations.

  • Larger-bodied bees (Xylocopa, Megachile) possess extensive foraging ranges, enabling them to traverse site boundaries into adjacent residential properties, where unmonitored commercial garden chemicals are applied.

  • Larger bees consume absolute greater volumes of nectar to meet metabolic flight demands, expanding their total cumulative dietary exposure to systemic contaminants.

• Assessment of Nesting Biology and Sociality:

  • Nesting substrate preference (ground-nesting vs. cavity-nesting) did not independently explain pesticide exposure variance:

  • Ground-nesting taxa (Agapostemon, Halictus, Lasioglossum) generally lacked detectable residues, but this relationship was confounded by small body size.

  • Melissodes, the largest ground-nesting genus evaluated, exhibited pesticide detection rates and concentrations comparable to cavity-nesting bees of similar mass (Megachile).

  • Social structure (solitary, communal, or eusocial) showed no significant main effect on pesticide detection rates or tissue concentrations.

• Toxicological Consequences of Detected Compounds:

  • Piperonyl Butoxide:

  • Function: Synergist that inhibits metabolic cytochrome P450 enzymes in insects.

  • Significance: Although not intrinsically toxic to bees, its presence blocks enzymatic detoxification, exponentially amplifying the lethal and sub-lethal potency of co-occurring insecticides (such as pyrethroids or neonicotinoids).

  • Source: Common additive in retail bug sprays and off-target exposure from domestic pet flea treatments.

  • Insecticides:

  • Fipronil (max $18.51\,ppb$ in Xylocopa): Induces motor and behavioral impairment, damages mushroom body brain architecture, and disrupts olfactory learning and memory acquisition.

  • Neonicotinoids (Dinotefuran max $2.74\,ppb$, Thiamethoxam max $1.31\,ppb$ in Melissodes): Known to disrupt spatial memory, reduce individual foraging efficiency, suppress colony growth, and impair wild bee reproductive output.

  • Fungicides and Herbicides:

  • Fungicides (4-Hydroxy-chlorothalonil max $22.25\,ppb$, Tebuconazole max $1.94\,ppb$) and herbicides (Diuron max $1.90\,ppb$, Metolachlor max $1.45\,ppb$) impair olfactory learning, shorten adult worker lifespan, alter gut microbiomes, and exhibit synergistic toxicity when combined with insecticides.

• Conservation and Policy Recommendations:

  • Urban green space networks (community gardens and restored grasslands) provide vital functional refuges that shield small-bodied native bees from agrochemical hazards.

  • Restricting cosmetic and non-agricultural pesticide applications across urban residential zones offers an effective mechanism to eliminate anthropogenic chemical stress on native pollinators without incurring agricultural yield losses.