The impact of family farming on Afrotropical flower fly communities (Diptera, Syrphidae): A case study in Tanzania

Abstract

To provide empirical evidence about the impact of family farming on Afrotropical flower fly communities (Diptera, Syrphidae), we established a large experimental setup in the Morogoro area (Eastern Central Tanzania) and quantified insect abundance and diversity in contrasting agricultural landscapes. Over the two years of this study, we collected 12,969 flower flies from 55 species and 3 subfamilies: Eristalinae (29 species), Microdontinae (2 species), and Syrphinae (24 species). The ten most abundant species contributed to 84.95% of specimens. Overall, we did not observe major changes in species richness or diversity between agroecological and conventional farming. In contrast, higher abundances of the two dominant species, Toxomerus floralis (Fabricius, 1798) and Paragus borbonicus Macquart, 1842 (69.49% of all specimens collected) were observed in agroecological treatments. This effect was more pronounced where the landscape features were more favourable to each of these species (i.e., in the plateau for T. floralis and in the mountains for P. borbonicus). Landscape provided a comparably much stronger effect than farming practices, and the percentage of variation explained by landscape, as a standalone factor, was approximately five times higher than for farming practices. Spatial heterogeneity and seasonality also provided a large and significant proportion of random variability. Our results stress how verifying a generally accepted paradigm of sustainable agriculture, "agroecology promotes abundance and diversity of beneficial insects", might require careful consideration, as, under field conditions, the impact of sustainable farming practices on insect communities might be embedded within complex, multi-layered ecological interactions.

Fig 1. Map of the experimental sites in Morogoro, Tanzania (see Acknowledgements for map copyright notice).

Fig 2. Total number of specimens for the ten most abundant flower fly species in the experimental setup.

Fig 3. Total flower fly abundances.

The bar plots illustrate the significant interactions: (A) Management x season (MPxSE) and (B) zone x season (ZOxSE). Significance letters for the pairwise tests are indicated (see Table 1).

Fig 4. Abundances of T. floralis and P. borbonicus.

The bar plots illustrate the significant interaction of management x zone (MPxZO) on abundance: (A) for T. floralis and (B) for P. borbonicus. Significance letters for the pairwise tests are indicated (see Table 1).

Fig 5. Shannon Diversity and evenness.

The bar plots illustrate the significant interactions of management x zone (MPxZO) on (A) Shannon diversity and (B) species Evenness. Significance letters for the pairwise tests are indicated (see Table 2).

Fig 6. Patterns of β diversity. Interaction of the effects of management and zone (MPxZO, see Table 2).

(A) Unconstrained and constrained ordinations (non-metric Multidimensional Scaling, and (B) Analysis of Principal Coordinates) of the abundances of 55 flower fly taxa in different zones (mountainous, plateau) and land management (agroecological, conventional). Results are shown for the whole dataset (1,440 samples) and for pooled replicates across sites (160 samples). Ellipses including 95% of samples are indicated.