Mapping potential malaria vector larval habitats for larval source management: Introduction to multi-model ensembling approaches

Abstract

Mosquito larval source management (LSM) is a viable supplement to the currently implemented first-line malaria control tools for use under certain conditions for malaria control and elimination. Implementation of larval source management requires a carefully designed strategy and effective planning. Identification and mapping of larval sources is a prerequisite. Ensemble modeling is increasingly used for prediction modeling, but it lacks standard procedures. We proposed a detailed framework to predict potential malaria vector larval habitats using ensemble modeling, which includes selection of models, ensembling method and predictors; evaluation of variable importance; prediction of potential larval habitats; and assessment of prediction uncertainty. The models were built and validated based on multi-site, multi-year field observations and climatic/environmental variables. Model performance was tested using independent multi-site, multi-year field observations. Overall, we found that the ensembled model predicted larval habitats with about 20% more accuracy than the average of the individual models ensembled. Key larval habitat predictors were elevation, geomorphon class, and precipitation 2 months prior. Mapped distributions of potential malaria vector larval habitats showed different prediction errors in different ecological settings. This is the first study to provide a detailed framework for the process of multi-model ensemble modeling. Mapping of potential habitats will be helpful in LSM planning. Authors summaryMosquito larval source management (LSM) is a viable supplement to the currently implemented first-line malaria control tools. Implementation of LSM requires a carefully designed strategy and effective planning. Identification and mapping of larval sources is a prerequisite. Ensemble modeling is increasingly used for prediction modeling, but it lacks standard procedures. We proposed a detailed framework for such a process, including selection of models, ensembling methods and predictors; evaluation of variable importance; and assessment of prediction uncertainty. We used predictions of potential malaria vector larval habitats as an example to demonstrate how the procedure works, specifically, we used multi-site multi-year field observations to build and validate the model, and model performance was further tested using independent multi-site multi-year field observations - this training-validation-testing is often missing from previous studies. The proposed ensemble modeling procedure provides a framework for similar biological studies.