Machine Learning Applications in Climate Change: A Science Mapping and Network Analysis

Stock image

Image taken from Pexels

Climate change is described as one of the greatest threats facing humanity, both in this current age as well as future generations. In addition to this, it is portrayed as a calamity that was caused by humans and poses severe threats to human health and safety and the environment in general. The National Aeronautics and Space Administration (NASA) defines climate change as a change in the typical meteorological or seasonal conditions of a particular location over the course of time. The combustion of fossil fuels, which in turn generates and emits greenhouse gases (GHG) like carbon dioxide and methane, is one example of an activity that can be either naturally occurring or caused by human activity (also known as anthropogenic). The buildup of greenhouse gases in the atmosphere of the planet causes a gradual warming of the atmosphere (which results in an increase in the average global temperature), which in turn causes changes in the weather and environmental conditions. On a similar note, the so-called greenhouse effect is caused by the decline in the earth's natural carbon reduction sinks that occurs as a result of the increasing concentration of greenhouse gases that are emitted.

Multiple pieces of research have indicated that climate change may lead to worldwide and unparalleled difficulties for civilisation. For instance, it is projected that the changes in weather and climate conditions that are brought about by climate change would bring about intense heat waves, acute droughts, and air pollution, in addition to the possibility of catastrophic flooding. These factors may lead to a lack of water supply, food insecurity, hunger, and malnutrition. In a similar vein, experts think that climate change may exacerbate the effects of pollution and environmental difficulties, both of which may provide significant dangers to human health, well-being, and safety. According to the World Health Organisation (WHO), climate change will have a significant influence on human health in a variety of other ways, including the disruption of drinking water supplies, the nutritional quality of food, and the availability of secure housing, amongst other possible consequences. According to global experts, drastic action is required to comprehend climate change, minimise its effects, and adapt to it in order to prevent its future catastrophic and costly implications. This is because climate change is already having an influence on humans, ecosystems, and the environment, and it is also having the capacity to do so in the future.

A great number of strategic policies and transformative initiatives have been offered over the course of the years in order to lessen emissions of carbon dioxide (CO2) and other greenhouse gases (GHG) as well as global surface temperatures in order to improve the ability to recognise and adapt to the effects of climate change. In a similar vein, sustainable methods and cutting-edge technology are currently being used in order to forecast and comprehend the effects of climate change on a variety of industrial sectors.  

A method that falls under this category is the application of computational techniques, which include artificial intelligence (AI), the internet of things (IoT), and deep learning or machine learning. Due to the unpredictability of climatic circumstances, it is necessary to have such sophisticated computational tools in order to effectively detect weather patterns at an early stage and estimate the effects of climate change. In light of this, the application of machine learning (ML) in the field of climate change research has emerged as a significant topic of investigation.

According to various environmental/climate reports, the global average temperature (GAT) has risen by about 0.85 ◦C (1.53 ◦F) between 1880 and 2020. The data shows that the planet GAT has increased by 0.08 ◦C (or 0.14 ◦F) every ten years since 1880, although the rate of warming has more than doubled per decade to 0.18 ◦C (0.32 ◦F) since 1981. Due to the changes in GAT, scientists predict that climate change could result in net damages with significant costs to humanity and the environment. According to some analysts, some regions of the world are already experiencing higher rainfalls, more floods, mudslides, rising ocean levels, loss of habitat, as well as severe heat waves. As a result, numerous analysts, policymakers, and scientists around the globe have prioritised the search for strategies, tools, and systems to increase the understanding of the global challenges immediately and potentially posed by climate change and global warming. The multidimensional nature of climate change has prompted multidisciplinary studies spanning different subject area.

The findings of this study suggest major studies on MLCC research have been focused on the impacts of climate change on agricultural production, weather conditions, and disaster prediction and preparedness, among others. The study by Crane-Droesch, examined the application of ML techniques in predicting the impact of CC on methods for corn crop yields in the United States. The impact assessment of crop yield was examined using a semi-parametric, and cross-sectional heterogeneous deep neural network approach. The results showed that CC has significant adverse effects on the yield of corn production, which could have long-term impacts on not only the environment but the socio-economic well-being of the farmers, regions, and global food chain.

Dr Samuel Mofoluwa Ajibade
School of Engineering and Technology
Email: [email protected]