The Impact of Climate Variability on the Spatial Dynamics of Malaria in the Global South: A Systematic Review of Modelling Studies, with a Focus on Vulnerable Populations
This systematic review of 39 modelling studies (2010–2025) reveals that temperature and rainfall are the primary climatic drivers of malaria transmission across the Global South, with machine-learning and hybrid models offering superior predictive capabilities while underscoring the critical need for integrated frameworks that address the heightened risks faced by vulnerable populations.
Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer
Malaria is a disease that thrives in the warmth and wetness of the tropics, carried by mosquitoes that breed in standing water and transmit parasites to humans. For decades, scientists have known that the weather plays a starring role in how this disease spreads. When temperatures rise, mosquitoes can survive in places that were once too cold for them, and when rains fall, they create new pools for breeding. However, the relationship is not a simple straight line; too much rain can wash away breeding sites, and extreme heat can kill the insects. As the global climate shifts, becoming more unpredictable with hotter days and erratic storms, public health officials face a difficult question: where will malaria appear next, and who will be most at risk? Answering this requires looking at the landscape not just as a map of countries, but as a complex system where weather, geography, and human life intersect.
To navigate this complexity, researchers have turned to computer models. These are not crystal balls, but rather sophisticated mathematical tools that take vast amounts of information—such as daily temperature readings, rainfall totals, and satellite images of vegetation—and use them to simulate how malaria might behave. Some models rely on traditional statistics to find patterns, while others use machine learning, a type of artificial intelligence that can spot subtle, non-linear connections that humans might miss. The goal is to move beyond guessing and toward predicting outbreaks before they happen, allowing health workers to send out mosquito nets and medicine to the right places at the right time.
A new systematic review brings together the best of this work to see what we have learned so far. Researchers gathered and analyzed 39 different studies published between 2010 and 2025 that focused on the Global South, a term encompassing the developing nations of Africa, Asia, and parts of Latin America where malaria is most common. The team looked specifically at how these studies used climate data to model the spread of the disease, paying close attention to whether they considered the people most likely to suffer, such as young children and pregnant women. By synthesizing these findings, the review offers a clear picture of how climate variability drives malaria and which tools are best suited to track it.
The analysis revealed that the vast majority of this research has taken place in Africa, with 26 of the 39 studies conducted on that continent. This focus reflects the heavy burden of the disease there, but it also highlights a gap in knowledge for other regions, as no studies from Latin America and the Caribbean met the criteria for inclusion. The remaining studies were spread across Asia and the Middle East. Across all these diverse locations, two climate factors stood out as the most consistent drivers of malaria transmission: temperature and rainfall. These two elements appeared in nearly every study, with temperature cited in over 90 percent of them and rainfall in more than three-quarters. While other factors like humidity, the greenness of vegetation, and elevation also played roles, the rhythm of the seasons—defined by heat and water—remained the primary engine of the disease's spread.
The researchers found that the way scientists model these relationships has evolved significantly over the last decade. Early studies relied heavily on mathematical and statistical models, which are excellent for explaining why something happens. These models are transparent and easy to interpret, helping policymakers understand the link between a rainy month and a spike in cases. However, the review noted that these traditional methods sometimes struggle with the messy, non-linear reality of nature, where a small change in temperature might have a huge effect, or where the impact of rain depends on how long it has been since the last storm.
In response, more recent studies have increasingly turned to machine learning. These computer algorithms are like powerful pattern-recognition engines that can digest massive datasets to find complex connections that simpler methods miss. The review found that these machine-learning approaches generally performed better at predicting exactly where and when outbreaks would occur, often achieving high levels of accuracy. Yet, they come with a trade-off: while they are excellent at prediction, they can be harder to understand, acting somewhat like a "black box" where the internal logic is not immediately clear to human observers. The most promising path forward appears to be a hybrid approach, combining the predictive power of machine learning with the explanatory clarity of traditional statistics. This blend allows scientists to forecast risks with high precision while still being able to explain the factors driving those risks to public health leaders.
Crucially, the review emphasized that climate does not act alone. The studies that focused on vulnerable populations—specifically children under five, pregnant women, and low-income rural communities—found that these groups bear a disproportionate share of the disease burden. For these populations, the risk is not just about the weather; it is about how the weather interacts with poverty. A rainy season might not lead to a major outbreak in a wealthy neighborhood with good drainage and screened windows, but in a rural village with poor housing and limited access to healthcare, the same rains can trigger a surge in infections. The research showed that models which ignored these social factors were less effective at protecting the people who needed it most.
The evidence gathered in this review suggests that while we have made significant strides in understanding the climate-malaria connection, there is still work to be done. Many models still rely on a limited set of weather data and often miss other important environmental details, such as soil type or the specific ways land is used by farmers. Furthermore, the social factors that make people vulnerable are frequently left out of the equations because data on them is hard to find or inconsistent. The authors conclude that the future of malaria control lies in building more integrated models. These next-generation tools would combine weather data, environmental details, and social information into a single, comprehensive picture. By doing so, they could help health systems become more resilient, ensuring that when the climate shifts, the protection for the world's most vulnerable people shifts with it.
Drowning in papers in your field?
Get daily digests of the most novel papers matching your research keywords — with technical summaries, in your language.