A recent study, published in The Lancet Regional Health – Europe, sheds light on the short-term effects of temperature and precipitation on the incidence of West Nile Neuroinvasive Disease (WNND) in Europe. Conducted by researchers from the E4Warning partners from the Barcelona Supercomputing Center (BSC) and international collaborators, the study utilized a nine-year dataset (2014–2022) to analyze 3,437 WNND cases across 20 European countries.

To estimate the association between lagged meteorological variables and WNND cases, the study employed a space-time-stratified case-crossover design. This design treats each case as its control, comparing exposures during the “hazard period” (the weeks before symptom onset) to exposures in comparable “reference periods.” Cases and controls were matched within the same region, calendar month, and year, as well as by the day of the week, enabling precise control of confounding factors such as geography, long-term trends, and seasonality.

For each case, weekly mean temperature and cumulative precipitation were attributed to the 8 weeks preceding the symptom onset day and matched control days. For instance, if a case occurred on May 17, 2019, in a given region, control days were selected from other Fridays in the same month and year. This method allowed researchers to infer associations between meteorological conditions and WNND incidence.

Meteorological data were obtained from the ERA5-Land climate reanalysis database, produced by the European Centre for Medium-Range Weather Forecasts (ECMWF). Researchers retrieved gridded daily mean temperature and daily cumulative precipitation data for the European continent over the entire study period (2014–2022). ERA5-Land provides atmospheric and land-surface variables with a spatial resolution of approximately 9 km × 9 km. To align with epidemiological data, daily meteorological data were spatially averaged at the NUTS3 level.

Key Findings:

• Temperature and Precipitation as Drivers: Weekly mean temperatures exceeding 25°C (the 90th percentile) were associated with 36.4% of WNND cases, with the strongest effect observed two weeks before symptom onset. Similarly, weekly cumulative precipitation above 40 mm contributed to 13.1% of cases, with a delayed peak effect three weeks before symptom onset.
• Regional Variation: The study highlighted consistent effects of temperature across regions but noted moderate variability in precipitation’s impact, reflecting differences in local ecosystems and mosquito ecology.
• Implications for Early Warning Systems: The findings underline the potential of integrating meteorological data into early warning systems. These systems could enhance preparedness for WNV outbreaks, particularly in areas lacking robust vector and host surveillance infrastructure.

Broader Implications:

This study reinforces the link between climate variability and the transmission of mosquito-borne diseases, emphasizing the need for a One Health approach to address the complex interactions between climate, ecosystems, and human health. The results provide actionable insights for tailoring early warning systems and vector control strategies to mitigate the growing burden of WNV outbreaks in Europe.

For more details, check out the paper: Moirano, G., Fletcher, C., Semenza, J. C., & Lowe, R. (2025). Short-term effect of temperature and precipitation on the incidence of West Nile Neuroinvasive Disease in Europe: a multi-country case-crossover analysis. The Lancet Regional Health – Europe, 48, 101149. https://doi.org/10.1016/j.lanepe.2024.101149

And in the E4Warning Zenodo repository: https://zenodo.org/records/14383572

As members of the E4Warning consortium, Gina Tsarouchi, Quillon Harpham (HR Wallingford), Lyra Tyson (UPF), and Remy Hoek Spaans (BSC) traveled to Sri Lanka in November to participate in a workshop organized by Dr. Lahiru Kodithuwakku and Gina Tsarouchi. The event, in collaboration with the National Dengue Control Unit (NDCU) and the Ministry of Health, focused on enhancing innovative tools for dengue prediction and control.

The workshop aimed to train participants in using the D-MOSS platform, a live dengue forecasting system combining satellite earth observation data, climate models, and statistical tools to provide probabilistic forecasts of dengue outbreaks. It also served as a valuable opportunity to adapt the methods to Sri Lanka’s unique public health context, informed by input from local professionals.

Bringing Together Experts from Across Sri Lanka

The workshop attracted over 80 participants from districts across the country, including entomologists, regional epidemiologists, medical officers of health, academics, NGOs, and representatives from the NDCU. This diverse group ensured that the discussions were enriched with practical insights and multidisciplinary expertise.

Dr. Preshila Samaraweera kicked off the workshop with an informative presentation on Sri Lanka’s progress in dengue control, highlighting current research projects and interventions. This set the stage for the D-MOSS training sessions, where participants had hands-on opportunities to test the system and provide feedback.

Training, Feedback, and Collaboration

The D-MOSS platform was well-received, with participants exploring its ability to visualize past, current, and forecasted dengue trends. One participant noted:

“These elements collectively offer a comprehensive view of past, present, and predicted dengue trends, supporting proactive planning and response.”

Following a lunch break, attendees learned about the statistical modeling methods used in E4Warning, with a focus on interpreting probabilistic forecasts. A group discussion then explored ways to enhance the dengue early warning system, including the integration of new data sources such as mobility patterns and socio-economic data. The day concluded with a presentation on the Mosquito Alert platform, sparking a lively conversation about how citizen science could contribute to vector control in Sri Lanka.

Lessons Learned and Next Steps

Reflecting on two intensive days of learning, several valuable lessons emerged from the experience. As highlighted during the workshop discussions, effective dengue control—particularly larval source management—requires a multi-sectoral approach. For example, Aedes mosquitoes often breed in artificial containers that fall under the responsibility of different sectors, such as road maintenance crews for potholes or port authorities for harbor areas.

Another critical area for improvement is the integration of entomological data into a digital infrastructure. Currently, much of this data is collected on paper, limiting its accessibility and use for coordinated control efforts. Digitizing these records would enhance monitoring, analysis, and intervention strategies.

Additionally, dengue prediction models could be further refined by incorporating mobility patterns and socio-economic census data, offering a more comprehensive understanding of how the disease spreads and where interventions could be most effective.

By continuously enhancing the digital tools developed through the E4Warning project, we aim to support the Sri Lankan National Dengue Control Unit (NDCU) in achieving its ambitious goals for dengue control and prevention.

E4Warning participated in the high-level conference “Research Perspectives on the Health Impacts of Climate Change,” organized by the Directorate-General for Research and Innovation (DG RTD) of the European Commission. This event, held in Brussels on February 19th and 20th, 2024, aimed to highlight the crucial nexus between climate change and global health issues.

Chloe Fletcher, a PhD candidate from the Global Health Resilience group at the Barcelona Supercomputing Center (BSC), attended the conference. Her research focuses on developing statistical models for predicting outbreaks of climate-sensitive infectious diseases, such as dengue and leptospirosis. Fletcher presented a poster titled “A Generalized Modelling Framework for Dengue Early Warning Systems in South and Southeast Asia,” a collaborative effort involving Ramy Hoek Spaans (BSC), Sophie Belman (BSC), Chloe Fletcher (BSC), Quillon Harpham (HRW), Gina Tsarouchi (HRW), and Rachel Lowe (BSC), all contributing to the E4Warning project.

This study aims to broaden the successful application of the Dengue Forecasting Model Satellite-based System (D-MOSS), originally implemented in Vietnam, to include Sri Lanka and Malaysia. By employing a Bayesian modeling approach and contrasting it with the conventional practices of dengue outbreak monitoring in each nation, the research seeks to identify which dengue predictors remain consistent or vary across these diverse regions.

The goal is to establish a universally applicable template for developing dengue early warning systems. The D-MOSS platform, which underpins this forecasting system, is managed by HR Wallingford. A key ambition of E4Warning is to collaborate intimately with local partners to deliver actionable data visualizations, thereby enabling an informed, data-driven approach to dengue management.

You can view the summaries of other posters presented during the conference at this link.