BLESSING MBALAKA
The recent snowfall is a picturesque spectacle that overwhelmed my WhatsApp stories. However, could there be a dark side to snow? Could it paint a picture of a worsening environmental crisis?
Could AI technology play a role in helping to mitigate the impacts of this worsening environmental crisis?
During a televised interview with eNCA, an associate professor of physical geography at Wits University, Jennifer Fitchett, said that, in isolation, a single weather event is seldomly attached as an indication of climate change. Instead, she mentioned that climate change is seen through long-term changes.
To briefly frame the problem at hand, the 2007 report by the IPCC (Interplanetary Council on Climate Change) noted that the overarching consensus from their findings suggests that human activity significantly contributed to climate change.
This same environmental change has been argued to contribute to the increase in environmental fluctuations. One such argument was raised by the Institute for Environmental Policy’s brief titled “Climate change and natural disasters: scientific evidence of a possible relation between recent natural disasters and climate change”.
The 2006 report found that GHG (greenhouse gas) emissions coincided with an increased rate of natural disasters. The human-induced 70% increase in hurricanes in Europe is one corroborating finding from this same report.
Resolving climate change is a time-sensitive dilemma. The dilemma requires urgency, and according to the 2022-2023 Caribbean/Latin America Disaster Readiness Manual, “If the planet warms by four degrees (as it is on course to do), an extreme heat event that would have occurred once in 50 years is 39 times more likely.”
The same manual also notes that if the temperature increases by 1.5%, “the same event is 8.6 times more likely.”
Predicting environmental phenomena has been a frustrating feature of meteorological studies for a long time. In 2020 alone, environmental disasters resulted in damage amounting to $22 Billion (about R389 billion) globally. Lenin Ndebele’s citation of Alize le Roux, a senior researcher for African Futures and Innovation at the Institute of Security Studies (ISS), noted that, since 1980, South Africa had lost nearly R640 billion to climate change.
The emergence of numerical weather prediction systems at the turn of the 20th century by the meteorologist Cleveland Abbe and physicist Vilhelm Bjerknes pioneered a call for the use of mathematical modelling and physics to enhance weather predictions. At the time of their proposal, atmospheric methods were weaker, and the immense benefits obtained from the invention of the computer did not exist.
Today, the use of computational methods in weather prediction has become a recurring feature of meteorological studies.
In the 100 years since the proposition by Abbe and Bjerknes, computers have emerged, evolved, and enabled meteorologists to simulate and predict weather conditions. However, the success rate of these predictions continues to limit confidence in these same systems.
This challenge can be noted by the Nasa-developed organisation SciJinks, which states that “A seven-day forecast can accurately predict the weather about 80% of the time, and a five-day forecast can accurately predict the weather approximately 90% of the time. However, a 10-day or longer forecast is only right about half the time.”
Computational evolutions, such as the age-old AI (artificial intelligence) technology, could potentially improve these forecasts. Artificial intelligence is a technology that dates back to the 1940s, yet its recent mainstream attention has opened the eyes of the public to the immense potential of the technology.
Can AI potentially help predict climate disasters? This enhancement will, of course, have to be an improvement on the current models.
According to a 2017-2020 report published by the weather intelligence service ForecastWatch, IBM, the industry standard for weather prediction technology, has already begun the use of artificial intelligence in its weather prediction systems, and they argue that their approach achieves this hope of enhanced accuracy.
The significance of these developments in AI is great because disaster risk practitioners and policymakers may be best prepared for climate risk management and strategies if they are able to utilise these tools to receive timely alerts.
Of course, sceptics will inquire about the accuracy of these AI models. These sceptical remarks are highly welcome because no prediction algorithm, to date, has a 100% success rate.
Dr Amy McGovern corroborates this in her interview on MyRadar Weather News. She highlighted how a slight variance in weather of one degree Fahrenheit (-17.2 degrees Celsius) could lead to freezing rain. By citing the butterfly effect, Dr McGovern warns that, at this stage, no model to date is 100% accurate. In the discipline of meteorology, a slight variable change can lead to unpredictable variations.
In weather forecasting, these minute differences, as noted in the butterfly effect, can have major implications for society and contribute to unexpected catastrophes. Simply put, the butterfly effect states that weather conditions between two weather systems can vary if there is a small variation at the start of a weather system’s formation.
The previously noted variance can lead to a totally different outcome between the two starting systems. AI, however, is a potential game changer for disaster risk management because any enhancement to the prediction of climate disasters could aid in the enhancement of preparedness.
Organisations that could benefit from this enhanced alert time include South Africa’s Gift of the Givers, The Association of Caribbean States, and the International Federation of Red Cross and Red Crescent Societies.
Climate change mitigation remains paramount, but tracking and monitoring through enhanced computational methods of AI could save lives and prevent some unexpected climate disasters.
The technology is still in its infancy, but the progress to be made is inspiring. This is immensely beneficial to the potential lives saved from enhanced response times and planning.
Blessing Mbalaka is a technology enthusiast and junior researcher at the Institute for Pan African Thought and Conversation, University of Johannesburg.
Daily News