Translate:

Login

CSIDNet Working Group on Early Warning Systems in LMICs: Progress Update

Written by the EWS working group

The CSIDNet Working Group on Early Warning Systems (EWS) in LMICs was established to foster collaboration, peer learning, and the practical application of existing climate-sensitive infectious disease models rather than developing another standalone tool. Our vision is to strengthen the transferability, interoperability, and operational use of early warning systems that support climate-informed public health decision-making.

Progress Update

Since its launch, the Working Group has refined its scope through technical discussions and a kick-off meeting. Members agreed to focus on evaluating, adapting, and testing existing early warning models in LMIC contexts, with particular attention to Africa. 

The group has reviewed major systems and modelling approaches, including WHO EWARS, ARBOTHAI, VECTRI, and other operational or research-based tools. These efforts aim to better understand model transferability, data requirements, technical limitations, and opportunities for operational implementation in resource-constrained settings. The ARBOTHAI framework itself demonstrates how climate information can be integrated into dengue forecasting to support operational public health decision-making. 

A key technical milestone has been the successful implementation of the VECTRI malaria model using CHIRPS rainfall data and ERA-Interim reanalysis data to simulate malaria transmission across Mali, Senegal, and Burkina Faso. This work generated operational indicators such as the Entomological Inoculation Rate (EIR), Human Biting Rate (HBR), vector density, larval biomass, and spatial risk maps that can directly support malaria early warning and preparedness. This work generated operational indicators such as the Entomological Inoculation Rate (EIR), Human Biting Rate (HBR), vector density, larval biomass, and spatial risk maps that can directly support malaria early warning and preparedness. 

To facilitate interpretation and dissemination of model outputs, the Working Group has also developed several interactive prototype dashboards that are currently under active development and continuous improvement:

These dashboards do not introduce new standalone models; rather, they build upon and extend existing community-developed modelling frameworks by making their outputs more accessible, interpretable, and actionable for end users. Specifically, we integrated outputs from established climate and disease models (e.g., VECTRI), together with climate reanalysis, seasonal forecasts, Earth observation products, and epidemiological datasets, into interactive decision-support platforms. Rather than developing new predictive models from scratch, our contribution focused on translating complex model outputs into intuitive dashboards that enable users to explore spatial risk patterns, temporal dynamics, and early warning indicators. This approach reflects the Working Group’s vision of strengthening existing community resources through integration, interoperability, and knowledge sharing. Importantly, applying an existing model to a new geographical context requires careful calibration and validation using local climate, environmental, epidemiological, and socio-demographic data. Without this adaptation, model outputs may not adequately represent local transmission dynamics or decision-making needs. Our work therefore demonstrates how established modelling frameworks can be responsibly adapted and operationalized for local contexts while preserving their scientific foundations and promoting reproducible, collaborative research.

In parallel, the Working Group has initiated the development of a collaborative review paper entitled “One Health Integration in Early Warning Systems for Climate-Sensitive Infectious Diseases.” The manuscript brings together expertise from across the Working Group and proposes an integrated One Health framework linking human, animal, and environmental surveillance, interoperable data systems, multi-hazard mapping, climate-informed risk indices, and operational early warning systems. This paper is intended to serve as one of the Working Group’s flagship scientific outputs.

Key Learnings

Several important lessons have emerged during the first months of the Working Group.

First, the principal challenge is not the lack of early warning models, but their limited interoperability, transferability, and operational implementation in LMICs.

Second, our experience has shown that practical testing and adaptation of existing models often provide greater value than developing new standalone tools from scratch. However, existing models should only be applied after evaluating whether they are appropriate for the local context, including the availability and quality of input data, disease epidemiology, climatic conditions, spatial and temporal scales, and the needs of end users. Pilot implementation helps identify where a model performs well, where calibration or adaptation is required, and where new approaches may be necessary. This process also reveals data gaps, technical constraints, and opportunities to improve operational readiness before wider deployment.

Finally, our experience within the Working Group has reinforced that effective climate-sensitive infectious disease early warning systems require sustained collaboration across climate science, epidemiology, veterinary medicine, environmental sciences, public health, data science, and decision-making institutions. Bringing together experts from these diverse fields has highlighted both the value of interdisciplinary collaboration and the challenges of integrating heterogeneous data, aligning methods and priorities, and translating scientific evidence into operational decision-making. Strengthening these collaborations remains a key priority for the Working Group.

How Others Can Get Involved

We warmly invite the CSIDNet community to explore the prototype dashboards and share feedback. If you think there are additional analyses, visualizations, or improvements that could strengthen the results, please let us know. Contact details are also available within each dashboard for anyone interested in providing feedback, discussing potential collaborations, or requesting additional information.

Next Steps

Over the coming months, the Working Group will: