Comparison of hand netting and pan trapping methods for estimating hover fly (Diptera: Syrphidae) diversity in the tropical agroecosystems of East Africa

Monitoring hover fly populations is essential for establishing baseline data on community dynamics and their potential links to ecosystem service provision. While pan traps and hand netting are widely used in monitoring insects in natural habitats, their comparative performance in estimating hover fly (Diptera: Syrphidae) diversity in Afrotropical agroecosystems remains understudied. This study assessed the effectiveness of yellow pan traps and hand netting over two consecutive years (2022–2023) in a cucurbit agroecosystem in Morogoro, Tanzania, across two contrasting landscapes (plateau and mountainous zones). Weekly sampling was conducted for eight consecutive weeks during both the rainy (April–June) and dry (September–November) seasons across 20 cucurbit plots. In each plot, nine yellow pan traps were deployed for 48 hours, and hand netting was conducted along three 15-m transects for 45 minutes per session. Hand netting consistently recorded higher hover fly diversity than pan traps across all Hill numbers (q = 0, 1, 2). Species accumulation curves reached asymptotes more rapidly with hand netting, indicating greater sampling efficiency and community coverage. Although more labour-intensive, hand netting provided a more comprehensive estimation of hoverfly diversity in both agroecological zones. In contrast, pan traps captured only a few additional low-abundance species overlooked by netting, which are unlikely to contribute significantly to ecosystem service provision. Therefore, hand netting proved to be the most robust and cost-effective method for monitoring hover flies in agricultural landscapes. These findings provided evidence to support the development of standardized monitoring protocols for hover flies in tropical agricultural landscapes.

Methods
Hover fly sampling was conducted using two complementary methods—hand netting and yellow pan traps—following modified protocols of Gervais et al. (2018) and Power et al. (2016). For hand netting, three permanent 15 m transects were set per subplot, 15 m apart and 7.5 m from plot edges. Sampling occurred weekly during cucurbit flowering, for 15 minutes per transect within a 2 m belt, on warm sunny days. Specimens were collected into ethanol-filled tubes. For passive sampling, three yellow pan traps (15 cm diameter) were randomly placed between transects to avoid spatial bias and capture natural microhabitat variability (Roulston et al., 2007; Westphal et al., 2008). Yellow traps were chosen to reflect the dominant flower colour of cucurbits and their proven attractiveness to pollinators (An et al., 2018; Farbenlernen et al., 2018). Traps containing 300 ml of soapy water were left for 48 hours. Collected hover flies were preserved in ethanol, identified morphologically under stereomicroscopes by S.K. and K.J., and voucher specimens were deposited at the Sokoine University of Agriculture (SUA) and the Royal Museum for Central Africa (RMCA), Belgium.