Background

This is a team project in the 2-weeks 2021 IBM Call for Code Research Challenge for Climate Change. Provided the powerful platforms and rich datasets from IBM, we attack the challenge theme: Climate Change.

Video

If your browser doesn't support this embedded video, please check here The introduction of our project starts at around 33:57.

Description

Thanks to the improvement of sanitary and hygiene, the worldwide burden of infectious disease has fallen over the past few decades. However, the covid-19 pandemic just shows us how great the threat to global health remains – especially as the climate crisis continues to affect the transmission of disease in a variety of ways. At the moment, changing environment is demolishing our planet’s defence system people used to rely on. One way in which climate change will affect the risk of diseases spreading is by making new areas into suitable homes for disease-carrying species. For example, rising temperatures and precipitation are making temperate, mountainous countries more susceptible to outbreaks of "tropical" or “low land” diseases like malaria. The arrival and rapid spread of the mosquito-borne viral disease across the world is one of the most significant challenge for public health developments of recent years. In order to prepare for future outbreaks, it is necessary to anticipate global regions that could become suitable for disease transmission.

In our project, by using the climate data from IBM PAIRS and The Weather Company, and surveillance of Anophelines’ occurrence from Malarial Mosquito Database , we aim to study the relationship between climate condition and survivability of disease vectors like Anophelines. More specifically, given a location’s climate conditions like temperature and precipitation, we trained a neural network (Fig 1.) to predict how possible Anophelines can survive in terms of climate condition. Due to time constraint, we trained our network using only data of Anopheline Funestus in the Sub-Saharan Africa, and the input features are monthly minimum temperatures and monthly maximum temperatures. And we applied our model to climate forecast during 2021-2040, the result is depicted by Fig 2. According to the result, due to effect of climate change, the climate condition of Portugal and Spain in the coming future is suitable for Anopheline to spread. And since there are travel between Sub-Saharan Africa and Portugal and Spain, the two countries will be in the risk of disease transmission.

Figure 1. Our simple model

Figure 2. Prediction result using data of 2022~2040

Acknowledgements

The website template was borrowed from Michaël Gharbi.