
Researchers in Google AI wrote an article called The Technology Behind our Recent Improvements in Flood Forecasting published on Thursday the 3rd of September, 2020.
A few things you might not know about floods:
- Flooding is the most common natural disaster on the planet.
- It affects the lives of hundreds of millions of people around the globe.
- Causes around $10 billion in damages each year.
Google has been working the previous years on flood forecasting.
They have more recently made efforts to improve flood forecasting in India and Bangladesh.
In doing so they are expanding coverage to more than 250 million people, and providing unprecedented lead time, accuracy and clarity.

They have devised a new way to approach this issue.
Namely, morphological inundation model: combining physics-based modeling with machine learning (ML) to create more accurate and scalable inundation models in real-world settings.
They combine this with alert-targeting model: "…allows identifying areas at risk of flooding at unprecedented scale using end-to-end machine learning models and data that is publicly available globally."
They have flood forecasting systems called HydroNets (presented at ICLR AI for Earth Sciences and EGU this year).
They describe a few important aspects:
- Forecasting Water Levels: whether a river is expected to flood.
"Once a river is predicted to reach flood level, the next step in generating actionable warnings is to convert the river level forecast into a prediction for how the floodplain will be affected."
- Morphological Inundation Modeling: "In prior work, we developed high quality elevation maps based on satellite imagery, and ran physics-based models to simulate water flow across these digital terrains."
According to Google AI doing this allowed warnings with unprecedented resolution and accuracy in data-scarce regions.
They mention that inundation modeling suffers from three significant challenges.
- "Due to the large areas involved and the resolution required for such models, they necessarily have high computational complexity.
- In addition, most global elevation maps don’t include riverbed bathymetry, which is important for accurate modeling.
- Finally, the errors in existing data, which may include gauge measurement errors, missing features in the elevation maps, and the like, need to be understood and corrected."
They compute modifications to the morphology of the elevation map that allow one to simulate the inundation using simple physical principles, such as those describing hydrostatic systems.
They train the model to use the data it is receiving to directly infer the inundation map in real time.

"The animation below illustrates the structure and flow of information in HydroNets…The output from this iteration of the network is then passed on to inform downstream models, and so on."

It is fascinating to see the progress made by Google AI.
This work may surely assist in saving millions of life.
This is #500daysofAI and you are reading article 469. I am writing one new article about or related to Artificial Intelligence every day for 500 days.