Spatial-temporal modelling and prediction of mosquito abundance across varying environmental gradients

In this unique work package, we will aim to combine both geostatistical and data science approaches for developing robust spatial-temporal models for predicting mosquito occurrence using environmental and socio-economic covariates and spatially referenced information on mosquito breeding sites collated from our real-time surveillance mobile phone application. Specifically, we will model occurrence using a range of socio-environmental determinants, such as temperature, rainfall, habitat, urban water infrastructure, housing quality and water sanitation. 

Aims and Objectives: 

The overarching aims are to harness the readily available data from the mobile phone application in conjunction with other open research data sources build models for predicting mosquito using spatial-temporal hierarchical Bayesian regression models which are inferred from Integrated Laplace Approximation (INLA). Other data of mosquitoes breeding behaviour (primarily the Aedes Aegypti, Aedes Albopictus, Culex and Anopheles species), life-cycle timings (egg time to hatching, extrinsic incubation period development time, adult lifespan), larval and adult feeding behaviour, as well as egg-laying habitat preferences, adult movement and activity times, will be collected to use as informative priors in our modelling framework, based on observations from community health workers. By working collaboratively with them, we will hope to achieve the following:

• We will use our models to investigate the relationship between mosquito abundance and habitat type in the different regions, by using mosquito abundance data collected at points along a gradient of human pressure from urban centres to more rural areas and incorporating habitat type as a fixed effect in the models.
• Generate electronic interactive maps to indicate areas which have significantly higher abundance, as where areas have exceedance probabilities exceeding the null thresholds to deem as high-risk.

Researchers in this group:

• Professor Wellington P. dos Santos
• Dr Anwar Musah
• Dr Ella Browning
• Clarisse Lins de Lima
• Ana Clara Gomes da Sliva