Forecasting Malaria in the Western Amazon

Reported cases of malaria in the western Amazon regions of Peru, Colombia and Ecuador have more than tripled since 2011. Responding to this epidemic has been challenging given large-scale environmental impacts and demographic changes combined with changing financial and political priorities. In Peru alone, malaria cases increased 5-fold since 2011. Reasons include changes in the Global Malaria Fund, massive flooding in 2012, the "mega" El Nino in 2016, and continued natural resource extraction via logging and mining. These challenges prompted the recent creation of the Malaria Cero program in 2017 with the goal to eradicate malaria by 2021. To assist in malaria eradiation, a team of investigators supported by NASA have been developing an Early Warning System for Malaria. The system leverages demographic, epidemiological, meteorological and land use/cover data to develop a four-component system that will improve detection of malaria across the western Amazon Basin. System components include a land data assimilation system (LDAS) to estimate past and future hydrological states and flux, a seasonal human population model to estimate population at risk and spatial connectivity to high risk transmission areas, a sub-regional statistical model to identify when and where observed malaria cases have exceeded those expected, and an Agent Based Model (ABM) to integrate human, environmental, and entomological transmission dynamics with potential strategies for control. Data include: daily case detection reports between 2000 and 2017 from all health posts in the region of Loreto in the northern Peruvian Amazon; LDAS outputs (precipitation, temperature, humidity, solar radiation) at a 1km and weekly scale; satellite-derived estimates of land cover; and human population size from census and health data. This presentation will provide an overview of components, focusing on how the system identifies an outbreak and plans for technology transfer.