Life-Saving Digital Oracles

It seems that 2023 is another year in which the “scare stories” about the apocalypse on the planet become reality. Pessimistic predictions are coming true before our eyes: abnormal snowfalls are replaced by the hottest heat wave on record. Fires are scorching acres of unique forests, floods are destroying crops and critical infrastructure, and earthquakes are killing tens of thousands of people.

For the past 12,000 years, our planet's landscape, climate and biodiversity have been shaped by the effects of the last Ice Age. Today, for the first time in history, humans are a key factor in the Earth's geology and ecosystem.

The burning of oil, coal and gas, uncontrolled deforestation, waste dumping, accelerated industrial development and unsustainable agriculture are increasing emissions of heat-trapping greenhouse gases. Under their influence, the average global temperature is 1° Celsius higher than it was 150 years ago. Weather is becoming more extreme, the oceans are warming and oxidising, sea levels are rising and glaciers are melting. The pace of climate change is accelerating, leading to an increase in the number of uncontrollable natural disasters.

According to the Centre for Research on the Epidemiology of Disasters (CRED), the number of natural disasters on the planet increased dramatically between 1980 and 1999 alone. During this period, there were 4212 disasters caused by natural hazards, killing some 1.19 million people, affecting 3.25 billion people and causing economic losses of USD 1.63 trillion.

Today, each of us has first-hand experience of the damage that extreme weather events and natural disasters can cause to countries and their citizens. These phenomena, especially when we are not prepared for them, often turn into real tragedies, causing social, environmental and economic damage, not to mention incredible human losses.

The good news is that scientific and technological advances are helping to greatly improve the accuracy of weather forecasts and provide early warning of natural disasters, saving thousands of lives around the world. Meteorologists have learned to collect, process and share large amounts of data more freely and quickly.

Machine learning and artificial intelligence have led to a quantum leap in global climate monitoring, modelling and forecasting, helping to make decisions that are critical to the safety and health of people around the world.

Early warning system 

On 26 December 2004, one of the most powerful earthquakes in human history occurred in the Indian Ocean. It measured 9.0 on the Richter scale. The tsunami triggered by the earthquake reached a height of 15 metres in some places. It killed between 230 and 280 thousand people in the region. This disaster prompted scientists to develop an updated concept for a global early warning system that could help prevent or minimise the devastating impact of natural hazards on communities around the world.

According to the developers, the system should synergize existing national and local forecast and warning capacities and expand efforts to better prepare for and mitigate catastrophes.

To be effective, early warning systems must be human-centred and include four basic elements: systematisation of the risks faced by technical monitoring and warning services; dissemination of meaningful warnings to those at risk; public awareness; and readiness to act.

Today, early warning technologies are available for almost all types of hazards and are operational in at least some parts of the world.

Life-saving warnings

Cambodia has implemented advanced technology for an inclusive early warning system. In case of danger, 130,000 citizens in the country receive audio warnings through the EWS1294 alert service, notifying local residents in advance of the need to prepare themselves and their families, as well as homes and businesses, for floods. Information preparation increases the psychological resilience of people and institutions to natural disasters.

The system, created more than a decade ago, used mobile technologies. However, the recent acceleration of the digital transformation of mobile infrastructures has significantly expanded the reach and capabilities of these systems.

Mobile penetration rates in the Asia-Pacific region indicate that 96% of the population already has access to mobile coverage, and another 430 million 5G connections will be installed by 2025, which will drive digital growth and technological innovation, and as a result, expand the opportunities for emergency alert systems including for remote areas.

Big data analysis

Earthquakes are among the most destructive natural disasters. At 04:17 on February 6, 2023, a series of catastrophic and deadly tremors hit southern and central Turkey and part of northern Syria. Large-scale destruction of residential areas caused the deaths of more than 50,000 people and injured 120,000.

For a long time, earthquake forecasting has remained an almost impossible goal for seismologists. However, Artificial Intelligence (AI) and the ability to collect and accumulate large amounts of data are revolutionizing the field.

Traditional methods of seismic data analysis involved manual interpretation of seismograms, which required enormous time resources and was not free from human error. Machine learning algorithms are able to quickly and accurately process large data sets, identifying patterns and correlations, and thus predict a future earthquake.

Researchers at Stanford University have developed a deep learning algorithm called ConvNetQuake. The algorithm was provided with data on more than 400,000 seismic events, which allowed it to learn to distinguish earthquake signals from background noise with phenomenal accuracy. During the tests, ConvNetQuake was able to detect earthquakes in 99% of cases, significantly outperforming previously known traditional methods.

The neural-based oracles

On March 1, 2011, an earthquake and the resulting tsunami caused a malfunction of the cooling system at the Fukushima Daiichi Nuclear Power Plant in Japan. The reactor cores melted at three power units at once and caused an explosion. This was the largest accident in the nuclear power industry since the 1986 Chornobyl disaster. More than 18,000 people died and nearly half a million lost their homes. Thousands of local residents cannot return home even decades after the accident because of the radioactive exposure threat.

The experience of the tragedy prompted researchers from the University of Tokyo to revolutionize the creation of neural networks to model complex processes in the earth's crust. The neural network model developed in Japan is capable of predicting the “behaviour” of fault systems responsible for most earthquakes. By inputting data on the geological properties of the fault system, the model can predict the way it will develop over time, providing a valuable and accurate forecast of the potential probability of future earthquakes.

The researchers set up the neural network to analyse and predict the behaviour of the complex fault system that triggered the 2011 earthquake. The result was an accurately reproduced and observable behaviour of the fault system. This gives hope that the neural network can become a valuable tool for predicting future earthquakes in similar regions.

***

One of the solutions that can mitigate the catastrophic effect of climate change is the transition to a carbon-neutral, green economy. For the European continent, which used to be dependent on Russian fossil fuels, diversification of energy sources is now a priority. Therefore, Ukraine, which before the Russian invasion was among the top 5 European countries in terms of solar energy development, can become one of the main strategic suppliers of electricity to Europe.

In the meantime, it is worth remembering that each of us can contribute to reducing emissions and ensure that we and our future generations of earthlings are able not only to survive but also to thrive on the planet, enjoying its breathtaking beauty and life-giving power.