How Guangzhou’s Early Warning System Contained Chikungunya as Climate Change Pushes Mosquito Diseases North

When a chikungunya fever outbreak ignited in Foshan, in China’s southern Guangdong Province, in July 2025, it took less than two weeks for the virus to cross into the neighboring megacity of Guangzhou. More than 4,800 cases were ultimately reported in Foshan, with 89 percent concentrated in a single district, and the epidemic peaked on 19 July with 681 new cases in one day. Yet all cases were mild, and no deaths occurred. A new study published in Frontiers in Earth Science argues that the difference was made not by any single breakthrough technology, but by an early warning monitoring system built methodically on the four pillars of public health surveillance: data sources, data analysis, early warning generation, and response protocols.

The research, led by a team from Sun Yat-Sen University’s School of Public Health and the Guangzhou Center for Disease Control and Prevention, takes an unusual approach for outbreak science. Rather than presenting new epidemiological data, the authors conducted a qualitative, single-case analysis of how Guangzhou’s Center for Infectious Disease Surveillance and Early Warning actually operated during the 2025 chikungunya response. Drawing on the center’s internal implementation plan, minutes from six thematic meetings held between July and October 2025, and direct observational notes, they mapped the city’s real-world practices onto the syndromic surveillance framework to identify both what worked and where the system strained.

The stakes of this analysis extend far beyond one Chinese city. Chikungunya, first identified in Tanzania in 1952, has now spread to 119 countries and regions worldwide according to the World Health Organization. Its vector, the same mosquitoes that carry dengue, is expanding its range as temperatures rise. Dengue incidence hit a record high in 2024, with more than 9.7 million cases reported in the Americas alone, more than double the 2023 figure. In Guangdong Province, dengue cases exceeded 10,000 in 2024, and the province’s experience shows that mosquito-borne disease translates directly into labor loss and economic burden. The Foshan-Guangzhou episode, the authors write, highlights the risk of the northward spread of tropical diseases driven by climate-widened mosquito habitats.

The scientific basis for this climate connection is well established. Temperature governs mosquito development rates, biting frequency, and the speed at which viruses replicate inside the insect, with optimal transmission typically occurring between 25 and 30 degrees Celsius. Precipitation creates the breeding habitats vectors depend on, though extreme rainfall can either promote or disrupt transmission depending on local hydrology. Studies in Guangdong have demonstrated that dengue incidence is significantly associated with temperature anomalies and cumulative rainfall, with lagged effects stretching up to eight weeks. These environmental drivers interact with vegetation cover and urban heat island effects to produce spatially patchy transmission risk, a pattern the researchers frame as part of Earth system dynamics shaping disease emergence.

Against this backdrop, Guangzhou’s system is notable for what it is not. Despite the growing literature on artificial intelligence in epidemic forecasting, the authors are explicit that no predictive AI, machine learning, natural language processing, or large language models were deployed during the 2025 response. The architecture rests on conventional information technology: a web-based national notifiable disease reporting system, a cloud-based data exchange platform for inter-agency coordination, and a short message service gateway for pushing alerts. The system’s strength lies in disciplined integration rather than algorithmic sophistication.

At the base layer, the system fuses multiple data streams. Domestic epidemic notifications, a global monitoring network, and local case data feed a real-time picture of transmission dynamics. Critically for a mosquito-borne disease, the system incorporates a meteorological data interface that continuously tracks air temperature, relative humidity, and cumulative rainfall, integrating these with landscape and urban ecological data. This coupling of environmental indicators with epidemiological surveillance allows the identification of spatiotemporal windows of elevated transmission risk, and the authors argue it lays a foundation for future integration with Earth observation systems and climate prediction models, even though the current application remains at the stage of indicator integration rather than quantitative prediction.

The analytical layer then processes this heterogeneous input through a risk categorization framework. Epidemiological, environmental, social, and even self-media information are collected and cleaned, with a web-based interface rapidly capturing early epidemic chatter from social platforms. A qualitative risk matrix, a standard tool in the Guangzhou CDC’s routine protocols, crosses three likelihood categories with three consequence-severity categories to yield a nine-cell grid that guides prioritization and response activation. The authors stress that this is a decision-support tool grounded in expert assessment, not a quantitatively validated predictive model. Virus characteristics, population mobility patterns, medical resource distribution, and community vulnerability all feed into the assessment.

Warning generation and response are where the system’s tempo becomes visible. Medical institutions must report suspected or confirmed chikungunya cases online within 24 hours, and public health emergencies must be reported through the emergency management system within 2 hours, with core case information pushed upward within 1 hour of a positive test. Once a warning triggers, a protocol the team calls the 1,335 mechanism kicks in: epidemic site disposal completed within 1 day, household surveys and adult mosquito elimination within 3 days, and vector density brought under control within 5 days. Screening strategies also shifted dynamically across three phases, from targeted nucleic acid testing of dengue-negative patients in districts ranked by importation risk, to citywide testing of all fever and unexplained joint-pain cases during local transmission, and finally to testing of defined suspected-case categories as the epidemic waned.

The study is candid about the system’s weaknesses. Cross-departmental data sharing lacks a mandatory mechanism, and incompatible formats among medical, customs, and environmental sectors caused delays that risked misjudging imported cases early in the epidemic. Manual entry errors in primary health institutions produced underreporting and misreporting, while social media data carried heavy noise. Perhaps most significantly, the linkage between meteorological parameters and actual transmission dynamics remains descriptive rather than predictive, leaving the system unable to anticipate climate-driven surges. The authors also flag institutional bottlenecks, including the recurring phenomenon of warnings being issued without corresponding action due to ambiguous responsibility boundaries, and the absence of a secure data desensitization mechanism to protect personal privacy.

The researchers propose four optimization directions: standardized data fusion with privacy-preserving computation to break down data barriers; a modular architecture that pairs national standardization with local adaptation; regional resource sharing through provincial cloud computing centers to ease financial pressure on poorer jurisdictions; and deeper integration of Earth system science, incorporating high-resolution climate projections, remote sensing vegetation indices, and land-use indicators into routine surveillance so that warnings become climate-informed and resources can be mobilized ahead of anticipated transmission peaks. They caution, however, that Guangzhou’s experience was enabled by robust infrastructure and computing power, and that its transferability to less digitized settings, whether other Chinese provinces or cities in the global south, remains an open question. As climate change continues to redraw the map of mosquito-borne disease, the study’s central lesson is that effective early warning depends less on futuristic algorithms than on the unglamorous work of connecting data, departments, and deadlines before an outbreak gets its foothold.

Subject of Research: Application of an early warning monitoring system for mosquito-borne disease control within the public health surveillance framework, using the 2025 Guangzhou chikungunya outbreak as a case study

Article Title: The application of the early warning monitoring system in mosquito-borne disease prevention and control within the public health surveillance system framework

Article References: Nie, Y., Zhen, R., Guo, Y., Ding, J., Zhang, W., Chen, Y., You, X., Yang, Q., Liu, W., & Hu, R. (2026). The application of the early warning monitoring system in mosquito-borne disease prevention and control within the public health surveillance system framework. Frontiers in Earth Science, 14, 1892700. https://doi.org/10.3389/feart.2026.1892700

Image Credits: AI Generated

DOI: 10.3389/feart.2026.1892700

Keywords: chikungunya, early warning monitoring system, public health surveillance, mosquito-borne disease, Guangzhou, climate change, dengue, syndromic surveillance, epidemic response, vector control, Guangzhou CDC, infectious disease