Use of climatic data and satellite imagery to model the abundance of Culicoides imicola, the vector of African horse sickness virus, in Morocco.

African horse sickness (AHS) is a vector-borne, infectious disease of equids caused by African horse sickness virus (AHSV). The only proven field vector of the virus is the biting midge Culicoides imicola. Following a recent epizootic (1989-91) of AHS in Morocco, light traps and automatic weather stations were operated for 2 years at twenty-two sites distributed over much of the country. The annually-averaged mean daily trap catch of C. imicola at these sites was negatively correlated with wind speed, and positively correlated with the average and mean annual minimum NDVI (Normalized Difference Vegetation Index, a remotely sensed measure of vegetation activity). There were no significant correlations between the mean daily trap catch and air temperature, soil temperature, relative humidity, saturation deficit, rainfall, altitude or the mean annual maximum or range of NDVI. The best two-variable model, which combined WindspeedMnAvMn (the average daily minimum wind speed of the least windy month) and NDVImin (the average annual minimum NDVI) as predictors, explained over 50% of the variance in the annually-averaged mean daily trap catch of C. imicola. There was a significant, positive correlation between minimum wind speed at night and the daily mortality rate of adult female C. imicola and it is suggested that the relationship between wind speed and the abundance of C. imicola arises from effects on adult mortality or dispersal. Considering several climatic variables, in North Africa NDVImin was most significantly correlated with total annual rainfall. It is suggested that the relationship between NDVImin and the abundance of C. imicola arises from the impact of soil moisture on both. It is proposed that areas of Morocco with higher levels of soil moisture in late summer or autumn provide more, larger and/or more enduring breeding sites for C. imicola, as well as supporting more photosynthetically active vegetation and hence having higher NDVI.