One of the most fascinating characteristics of artificial intelligence is near-boundless potential. It is a technology that learns by itself, enabling it to adopt numerous applications.
Nowadays, people recognize AI as text generators like ChatGPT. However, Google shows that this technology can do so much more for humanity.
READ: AI excels in weather forecasting
Google Research announced it developed a highly advanced artificial intelligence that can predict floods seven days in advance. More importantly, it works with limited data, allowing third-world countries to use it.
How does this Google AI work?
Researchers published their findings in Nature in “Global prediction of extreme floods in ungauged watersheds.” They call this Google AI the Flood Hub tool.
Google Research explains most hydrology models from national and international agencies are state-space models based on these data:
- Daily inputs like precipitation and temperature
- The current state of the system, such as soil moisture
On the other hand, FloodHub uses Long Short-Term Memory (LSTM) networks, which can make time-series predictions. In other words, it can make a series of predictions in a sequence of preset periods.
It processes input data like current weather conditions to produce updated state information and output values. Specifically, it uses these data from governmental sources:
- Static watershed attributes representing geographical and geophysical variables
- Historical meteorological time-series data
- Forecasted meteorological time series over a seven-day forecast horizon
As a result, Flood Hub can predict floods seven days in advance with limited information.
More importantly, this Google AI can predict extreme weather events more accurately than GloFAS version 4.
The Global Flood Awareness System is the current state-of-the-art global flood forecasting system.