Comparison of Mosquito-Magnet trap and livestock-aspirated collection of mosquitoes and the implications for modelling mosquito-borne animal diseases

Trap-collected mosquito surveillance data are widely used for modelling transmission of mosquito-borne animal diseases. Several input parameters are estimated from such data such as vector-to-host ratios, between-host attraction and biting rates. This assumes that traps are adequate surrogate hosts and the numbers and species of mosquitoes they attract are representative of those attracted by animal hosts. We evaluated the qualitative and quantitative differences in mosquito collections obtained by aspiration from cows and sheep and simultaneously by capture via Mosquito-Magnet (MM) traps set on a dairy farm. Several different mosquito species that landed and at times fed on livestock hosts were not captured simultaneously by the MM trap. Furthermore, orders of magnitude more mosquitoes were captured by aspiration on the livestock hosts compared to the MM trap. The proportion of blood-fed mosquitoes found in the MM trap was almost negligible (1.7%), while 41% and 14% of mosquitoes collected from the cow and sheep were blood-fed, respectively. This reveals the flaw of using mosquito surveillance data - primarily collected to gain insight into presence or absence of mosquito species and geographical distribution - for mosquito-borne animal disease transmission modelling.