Flooding has the potential to be a devastating event, with the potential to take lives, homes and livelihoods away from those affected. Furthermore, the threat posed by flooding on a global level appears to be increasing.
It has been reported by the World Bank that over 1.81 billion humans, representing more than one fifth of the entire globe's population, are now in danger of being exposed to "substantial flood risk".
Many other studies are cautioning that climate change is leading to more frequent and powerful floods. Recently, the United Nations issued a report suggesting that the danger of coastal floods could be five times higher this century.
In order to limit the effect of flooding, a selection of flood tracking companies have begun utilizing artificial intelligence (AI) to alert people before the levels of water increase.
Situated in Bergen, Norway, 7Analytics combines computer scientists and geologists to create real-time flood predictions for corporations and local governments. This way, they can identify those areas that are more prone to flooding and damage.
According to Jonas Torland, co-founder of 7Analytics, it is possible to predict flooding occurrences up to one week in advance based on weather predictions down to a very specific level.
This information takes into consideration the direction and impact of the water flow. We predict that in five days your entrance will be inundated with 50cm of water, and also when it will begin to recede.
This firm's AI software conducts an in-depth analysis of the weather conditions, geography of the land and rivers, degree of development in an area, and the drainage capacity. Prior to this, these flood prediction computations were managed using standard programing, but now, with the use of AI featuring more powerful computing and its aptitude for learning, the job can be done at a much faster rate with less need for human participation.
Mr Torland's company was established in 2020 and he says that it was challenging for them to obtain investments from backers at first. Nonetheless, he mentions that now in under 3 years, awareness has gone up that flooding is an issue that is getting more intense, and that has brought a shift in the situation.
Mr Torland commented that the situation has changed drastically and that investors are approaching them as they have come to recognize the merits of being proactive about flood prevention.
London-based company Neara utilises AI in order to generate digital flood modelling for the electricity industry. This modelling allows for the prediction of flooding, so that electricity networks can take the necessary steps to decrease the potential destruction.
It asserts that its information can demonstrate where it would be secure to restore electrical power after a flood, which wires should be disconnected, to which places power should be diverted, and the exact spots that engineers ought to be dispatched to.
Mary Cleary, chief marketing officer at Neara, believes that electricity loss is one of the most destructive issues that could arise from unpredictable flooding.
Ms Cleary comments that the digital models which it fabricates for the utility networks "depict with extraordinary precision every single subtleties there are, from the exact breadth of different wires, to what the adjacent constructions and vegetation appear to be like".
7Analytics and Neara both require payment for their service, while Google recently launched its no-cost Flood Hub platform to present flood warnings in over eighty nations.
This was first released in India in 2018, and then expanded in 2022. This year, it was introduced to both the United Kingdom and the United States.
Flood Hub utilizes satellite imagery to assemble digital depictions of rivers and the adjacent regions. AI software is subsequently employed to simulate the likelihood of flooding from intense precipitation and how and when it could occur.
It is said to provide warning periods of anywhere between two and seven days, thereby allowing authorities and people to prepare for evacuations or take other precautionary measures.
In India, Flood Hub now provides 45 million people with flood alerts through text messages, while in Chile, TV stations are relaying the warnings.
Yossi Matias, Google Vice President of Engineering and Research and Crisis Response Lead, emphasizes the severity of floods as one of the most destructive natural disasters, impacting millions of people annually. He highly emphasizes the importance of warning people before floods actually strike. So it’s an exciting development in terms of making predictions in these data-limited environments.”
He states that they have discovered AI technology can generate forecasts even when limited historical data is available. "These kinds of wide-reaching models have the capacity to learn from occurrences in other rivers with more data and history. It's a great progress in relation to making forecasts even in these data-poor circumstances."
That has allowed us to make forecasts in 23 African countries where we have fewer historical records.
Investigate further into AI through reading extra stories
Nevertheless, weather and computer professionals caution that when dealing with the data on which the AI relies, one must be mindful.
According to Dr Amy McGovern, a computer scientist at the University of Oklahoma School of Meteorology, the quality of the models is dependent on the information they acquire.
An instance where this might be the case is an area that infrequently experiences flooding. With limited data available, it may be more difficult for the flooding in those places to be accurately predicted.
Dr McGovern, researching the credibility of AI in forecasting the weather, remarks: "In order to build a successful AI model, you need to have reliable data. We are actively working on addressing the issue of bias too."
The cornerstone of reliability is ensuring conscientiousness and integrity while constructing artificial intelligence, as well as taking heed of any prejudices present in the data.
Despite the fact that it is a laborious task, she states AI can alleviate much of the burden, as it has ample computing power that can complete the task more quickly than conventional computing.
Mr Matias is hopeful about the potential AI can bring about regarding flood forecast. He states, "My aim would be to see that no one is taken off guard by a flood, or any other natural calamity, such that individuals are able to be ready and we don't have to endure fatalities or destruction in places that must not."
Comments