Temperature acts on malaria transmission like an arch. The window runs from roughly 16C to 34C [2], eighteen degrees wide [18], and a district's position inside it decides the sign of whatever warming arrives. Highlands that sat below the lower bound climb toward the productive middle; lowlands already near the top cross out of it, and neither the mosquito's life cycle nor the parasite's development keeps pace. That is how one century of warming shows up as better transmission conditions in East and southern Africa [8] and as suppressed transmission in West and central Africa [9].
The continental total is small partly because of how the arithmetic works. About one excess case per 1,000 children since 1901 [7] is a tenth of a percentage point of prevalence [19], and it is the sum of two regional numbers pointing in opposite directions. Tamma Carleton, an environmental economist at UC Berkeley and a coauthor, told Eos the finding is not about net aggregate effects across a large and diverse continent but about a burden shifting across space [13].
What the design buys is the word "caused." The team ran its comparison against a hypothetical world without human-caused climate change [11], the machinery built for attributing individual weather events, which Carleton notes is rarely pointed at public health outcomes [12]. That gives a stated counterfactual rather than an assumed one, which is more than a temperature-prevalence correlation delivers. The counterfactual is still a model output, and the measured outcome is prevalence in children aged 2 to 10 [3], not severity and not deaths.
The thing the effect size does not tell you is where money should sit. Carleton calls the changes modest next to the reductions already achieved by eradication campaigns alongside economic development [14]. Erin Mordecai of Stanford, who was not involved, reads the increase claim as nuanced: possibly subtly true in places like southern and highland East Africa where temperatures were previously sub-optimal, and small even there against drivers such as control programs [15]. Both readings hold, since climate here is a slow term added to a fast one.
Mordecai's other remark is the durable one. The thermal biology was predicted first and is now validated at continental scale [16]. A validated response curve is a forecasting instrument in a way a single continental effect size is not: given a district's temperature now and its trajectory, you can say which side of the arch it is walking toward. The authors expect the pattern to amplify as warming continues [6].
So, a view with its conditions attached. If the amplification holds, a program that ranks districts for nets, spraying and surveillance by prevalence over the past decade is working from a map that ages faster each year. Nothing in the paper allocates a budget, and the older dispute over whether climate matters at all for African malaria [10] is answered here only as yes, slightly, and unevenly.
