20 Mitigating the threat of insecticide resistance in African malaria vector 21 populations requires comprehensive information about where resistance occurs, 22to what degree, and how this has changed over time. Estimating these trends is 23 complicated by the sparse, heterogeneous distribution of observations of 24 resistance phenotypes in field populations. We use 6423 observations of the 25 prevalence of resistance to the most important vector control insecticides to 26 inform a Bayesian geostatistical ensemble modelling approach, generating fine-27 scale predictive maps of resistance phenotypes in mosquitoes from the 28Anopheles gambiae complex across Africa. Our models are informed by a suite of 29 111 predictor variables describing potential drivers of selection for resistance. 30Our maps show alarming increases in the prevalence of resistance to pyrethroids 31 and DDT across Sub-Saharan Africa from 2005-2017 as well as substantial 32 spatial variation in resistance trends. 33 34