[A-CG34] Climate Variability and Predictability on Subseasonal to Multidecadal Timescales

[ACG34-10] Predicting malaria outbreaks from sea surface temperature and sea ice concentration variability with machine learning

*Patrick Martineau¹、Swadhin Behera¹、Masami Nonaka¹、Venkata Ratnam Jayanthi¹、Takayoshi Ikeda²、Noboru Minakawa³、Philip Kruger⁴、Qavanisi Mabunda⁴ (1.Japan Agency for Marine-Earth Science and Technology、2.Blue Earth Security Co., Ltd., Tokyo, Japan、3.Nagasaki University, Institute of Tropical Medicine, Nagasaki, Japan、4.Malaria Control Programme, Limpopo Department of Health, Tzaneen, South Africa)

キーワード：Malaria prediction, Machine learning, Climate variability

Malaria is the cause of substantial socioeconomic burden for African countries. The planning of prevention interventions, that can effectively reduce this burden by mitigating the spread of malaria parasites by the anopheles mosquitoes, can benefit from early-warning predictions of malaria infection rates. Recent studies have demonstrated that there exists a statistical connection between major modes of tropical climate variability and malaria outbreaks, which could potentially contribute to producing skillful early warning predictions. In this work, we demonstrate for the first time that, in addition to tropical variability, Antarctic sea ice can be used to produce skillful predictions several seasons ahead. Sea ice concentration is especially useful to predict malaria incidence at the beginning of the high-risk season (September to December), a period for which tropical modes of climate variability are inadequate predictors. We find that the sea ice-malaria connection is established through an atmospheric Rossby wave train propagating around Antarctica, towards South Africa, where it induces high-pressure anomalies that contribute to enhancing atmospheric moisture fluxes from the Mozambique channel. As a result, the enhanced precipitation accumulated over the months leading to the outbreak contributes to increasing soil moisture, which likely modulates in a favorable way mosquito breeding sites.