The Dengue MOsquito Simulation from Satellites (D-MOSS) project is developing a dengue fever early warning system. The tool generates dengue warnings for likely outbreaks several months in advance. One of the key components is a water assessment module that provides the additional benefit of improving water management in transboundary river basins.
- Prediction of dengue outbreaks issued every month, up to seven months in advance at a district level, covering an entire country.
- Seasonal forecasts (six months) of water availability, covering the whole of the country as a catchment scale.
- Graphical and tabular portrayal of probability of future dengue outbreaks and water availability indicators presented in both English and the local language.
- Supporting information on recommended actions to be taken, provided by the decision makers, such as the Ministry of Health and
- Ministry of Natural Resources and Environment, and based on the forecasts and associated uncertainty.
Key technical features
- Incorporation of a variety of Earth observation (EO) data products ranging from historic observations to the latest state-of-the-art Sentinel missions.
- Hydro-meteorological and societal parameters are analysed in order to infer dengue fever outbreaks.
- Uses a common spatio-temporal analysis ‘grid’ with a ‘Polygon Series’ structure to integrate historical stressor datasets with each other and with historic dengue fever incidents, which are then input into a statistical model which provides forecasts based on future seasonal forecasts of these stressors.
- Accesses the UK Met Office seasonal forecasts (six-months lead time), a series of freely available optical and radar EO data from GPM (NASA/ JAXA), Sentinel and SMOS (ESA), SMAP, Aqua and Terra (NASA) such as rainfall, temperature, humidity, soil moisture, land use, population density. This is used to predict dengue outbreaks and water availability.
- The spatial resolution of the EO data used ranges from tens of metres to kilometres, depending on the application. For land surface temperature the Sentinel-3 spatial resolution is 300 metre, for soil moisture Sentinel-1 has a spatial resolution of 10 metre, for rainfall GPM’s resolution is 10 kilometre.
- A new forecast for each district is issued every month and during outbreaks the forecast is issued on a weekly basis.
- The architecture of the solution relies on open and non-proprietary software and on flexible deployment into platforms including cloud-based virtual storage and application processing.
- Web-based portrayal system and numerical model forecasting engine deployed on commercial cloud services which can be accessed via desktop/laptop devices with an internet connection. Architecture includes PostGIS spatially enabled relational database and Linux.
- Open-source solutions are employed where possible together with widely known development languages and tools.
- Built in capability to replicate the generic design in other parts of the world and for other diseases.
The D-MOSS dengue forecasting system is being implemented in Vietnam, supported by IPP. Considering the current trends in dengue epidemics, the D-MOSS dengue forecasting system will assist the Vietnamese Ministry of Health to put timely dengue control and prevention measures in place. The project will also estimate the likelihood and severity of future dengue outbreaks under a range of climate change, land-use and water management scenarios. This will allow the Vietnamese Ministry of Health to formulate appropriate interventions and policies.
HR Wallingford with headquarters in Wallingford in the UK, is a not-for-profit, independent research organisation, formed in 1947. HR Wallingford specialises in research and consultancy which includes early warning systems, water resources and climate change adaptation. HR Wallingford works closely with government and non- governmental organisations worldwide to provide the evidence-based tools to support decision making related to water availability and climate change adaptation across a range of sectors.