Mosquito community composition in South Africa and some neighboring countries

Abstract

Background: A century of studies have described particular aspects of relatively few mosquito species in southern Africa, mostly those species involved with disease transmission, specifically malaria and arboviruses. Patterns of community composition such as mosquito abundance and species diversity are often useful measures for medical entomologists to guide broader insights and projections regarding disease dynamics and potential introduction, spread or maintenance of globally spreading pathogens. However, little research has addressed these indicators in southern Africa.

Results: We collected 7882 mosquitoes from net and light traps at 11 localities comprising 66 species in 8 genera. We collected an additional 8 species using supplementary collection techniques such as larval sampling, sweep-netting and indoor pyrethrum knockdown catches. Highest diversity and species richness was found in the Okavango Delta of Botswana and in South Africa's Kruger National Park, while the lowest diversity and abundances were in the extreme southern tip of South Africa and in semi-desert Kalahari close to the South Africa border with Botswana. Species composition was more similar between proximal localities than distant ones (Linear model P-value = 0.005). Multiple arbovirus vector species were detected in all localities we surveyed (proportion of vector mosquito numbers were > 0.5 in all locations except Shingwedzi). Their proportions were highest (> 90%) in Vilankulo and Kogelberg.

Conclusions: Multiple known arbovirus vector species were found in all study sites, whereas anopheline human malaria vector species in only some sites. The combination of net traps and light traps effectively sampled mosquito species attracted to carbon-dioxide or light, accounting for 89% of the 74 species collected. The 11% remaining species were collected using supplementary collection techniques mentioned above. The diversity of species weas highest in savanna type habitats, whereas low diversities were found in the drier Kalahari sands regions and the southern Cape fynbos regions.

Keywords: Arboviruses; Diversity measures; Malaria; Mosquito community composition; Mosquitoes; Shannon index; Vectors.

Fig. 1

Map of collection sites and their land cover type within a 3 km radius area (a-k). The latest (2013) University of Maryland Land (UMD) cover type data was used. Snow/Ice and Barren land cover type are omitted from UMD legend as our study localities do not include any of those categories. Capital letters superimposed on the map insert correspond to equivalent small letter landcover charts

Fig. 2

Traps used for this study. a Mosquito net trap baited with polystyrene box containing dry ice placed on the ground in centre of trap. b Center for Disease Control (CDC) light trap with white light baited with dry ice

Fig. 3

Mosquito species composition pie charts of combined net and CDC-light traps. a Vilankulu, MZ. b Moremi, BW. c Khwai, BW. d Shingwedzi, KNP, SA. e Tshokwane, KNP, SA. f Skukuza, KNP, SA. g Lower Sabie, KNP, SA. h Lapalala, Limpopo, SA. i Tswalu, Northern Cape, SA. j Rooiport, Northern Cape, SA. k Kokelberg, Western Cape, SA. Species with densities below 2% were grouped into respective genera as spp. Genus name abbreviations are consistent with [95]. Dark tan color was used for Aedomyia species; shades of greens for Aedes; shades of red and oranges for Anopheles; shades of yellow for Coquellittidia; shades of blue for Culex; shades of purple for Mansonia; olive green for Mimomyia; and lime green for Uranotaenia. Numbers in outer pie slices indicate proportions of individuals from total net and light trap catches within each locality. Inner pie shows the relative proportions of each genus

Fig. 4

Species diversity measures for each location. Shannon’s index (top X axis) are represented in gray bars, the total number of mosquito specimens collected (bottom X axis) in black, the number of unique genera (top axis) in light blue and the total number of unique species (bottom X axis) are in dark blue

Fig. 5

Species composition similarity. a Dendrogram based on the pair-wise Morisita-Horn Index, a measure of species composition similarity between any two localities. b Relationship between species composition similarity measured as Morisita-Horn Index and geographical distance (Linear model slope = -0.0002, R² = 0.1399, P-value = 0.005)