|Paper title||Using EO parameters to model Vector Diseases and their drivers|
|Form of presentation||Poster|
The H2020 Mood project is focused on using ‘big data’ to provide risk assessments to Public Health professionals for a range of diseases including many Vector Borne Diseases such as Dengue, Tick Borne Encephalitis, West Nile Fever, and Congo Crimean Hemorrhagic Fever. Producing these risk assessments requires substantial support to provide the covariate datasets that drive the disease models. These datasets span a wide range of types of information, ranging from demographic, socio-economic and agricultural information to Earth Observation (EO) imagery.
This last set of environmental and climatic parameters are derived from a variety of EO and terrestrial sources , and comprise both near real time and long-term synoptic summaries produced by data reduction techniques such as Temporal Fourier Analysis of extensive time series of decadal imagery. The distributions of both disease vectors and hosts are also produced as additional drivers, either using species aggregation protocols, or estimated by spatial distribution modelling employing machine learning techniques and ensembling the outputs of several methods.
The project includes dedicated work packages to identify source, acquire, process, and supply the covariate datasets that drive the disease models to the modelers within the project as well as to external users that have requested access to such data. The presentation provides an overview of these data streams and driver modelling modelling techniques, taking examples from a number of diseases. These procedures include not only the driver modelling, but also parameter identification and selection, variable data reduction, static and dynamic suitability definition and masking, and data dissemination.
The presentation also discusses the particular requirements of providing such information to both the modellers, and to external users in the Public health arena who often need the products produced by academic research teams as publishable outputs to be adapted and adjusted for more practically oriented use, as well as made accessible in formats more suited to risk assessments than raster images.