A global survey of Flood Early Warning Systems, conducted
by the UN University’s Canadian-based Institute for Water
Environment Health, revealed that the majority of flood
forecasting centres in flood-prone countries lack critical
resources to carry out their functions. They have designed
new tools, namely: a Flood Mapping Tool. The new tools
will help to create inundation and flood risk maps and
bypass the high-cost issue by using big Earth data,
artificial intelligence models, open data, and cloud
About 90% of natural disasters are water-related – most
notably cyclones, floods and droughts.
Since 2000, over 5,300 water-related disasters have been
reported across the world, with over 325,000 fatalities
and economic losses exceeding US$1.7 trillion. Floods
alone account for approximately 54% of all water-related
In 2020, floods in South Asia affected more than 17.5
million people and caused more than 1,000 deaths. The
economic losses are still being calculated and are
expected to be in billions of US dollars. A similar
trend is observed in East Africa, where nearly 6 million
people were affected in 2020, with 1.5 million of
them forced from their homes.
Forecasts are key to mitigating the worst effects of
floods. However, a global survey of Flood Early Warning
Systems, conducted by the UN University’s Canadian-based
Institute for Water Environment Health, revealed that
the majority of flood forecasting centres in flood-prone
countries lack critical resources to carry out their
The centres lack the ability to improve the spatial
coverage and resolution of early warning systems.
They also can’t generate historical inundation and
flood risk maps. An inundation map points out the
specific area flooded by a particular flood event.
Inundation maps are critical for building flood
risk maps. They are created either by using hydrodynamic
or empirical models. The hydrodynamic models are
data-intensive as they need people to be on the ground,
gathering data from river and stream gauges.
Empirical models depend on remotely sensed data.
Unfortunately, the Global South lacks the infrastructure
and data to calibrate and run hydrodynamic models.
Today, gauging stations in North America outnumber
those in the 20 most water-stressed countries by
more than 10 to 1.
Developing inundations maps at the national level using
conventional techniques is a costly exercise. In
Canada, for example, it is expected to take a decade
and US$350 million to update national inundation maps.
Existing inundation and flood risk maps in most developing
countries are out-of-date. They also don’t consider
rapid urban development or the impacts of climate change.
Designing flood-risk maps
I’m part of the UN University’s Canadian-based Institute
for Water Environment Health team closing this knowledge
gap. We’ve designed new tools, namely: a Flood Mapping
Tool, which has just been released, and a Flood Risk
Prediction tool, which we’ll launch next year.
Our new tools will help to create inundation and flood
risk maps and bypass the high-cost issue by using big
Earth data, artificial intelligence models, open data,
and cloud computing.
They will provide critical input to flood mitigation
and emergency response, land use planning and investment
in resilient infrastructure, insurance schemes, and
overall public awareness of flood risks.
Other, similar, online tools include a platform launched
by the International Water Management Institute that
maps significant floods in South Asia from 1980 to 2011.
And the European Commission Joint Research Centre
launched an online tool in 2016 which provides free
access to global surface water indices.
Our flood mapping tool builds upon these initiatives
and improves the spatial and temporal resolution of
the inundation maps. We also focus on the Global South
as a whole, where the data and information gaps are
prominent and annual losses due to floods are high.
There are three actions that governments can take
to lessen the damage of future floods:
1)Improve the accuracy of flood maps to capture the
true extent of historical floods, rivers, streams,
and water bodies.
2)Improve the risk mapping and introduce policies to reduce risk and
3)Invest in the infrastructure to reduce risk at the community level.
The tools we are creating will help to provide
the critical data and evidence to implement these measures.
The Flood Mapping Tool, is the first to be released
as part of the UN University’s Canadian-based Institute
for Water Environment Health’s Web-based Spatial
Decision Support System. The aim is to address
information gaps in flood early warning and risk
This Tool generates inundation maps for significant
floods from 1984 to the present using publicly available
data on Google Earth Engine. It relies on a “data
cube” – spatially overlapped pixels of Landsat satellite
imagery captured over a period of time. This eventually
reveals inundation patterns over space and time.
By doing this, it allows the impacts of inundation
on various socio-economic sectors – such as agriculture,
forestry, transportation and communities – to be
During the Flood Mapping Tool’s two-year design,
review and testing process, our team engaged an
extensive network of water-related disaster experts
and representatives from disaster management agencies
from a range of countries. These included Afghanistan,
Bangladesh, Bhutan, India, Pakistan, Nepal, and Sri Lanka.
Google and MapBox also supported the tool’s development
through their research and education programmes. The
development team also worked with experts at the
Asian Disaster Preparedness Center, Thailand, and
McMaster University, Canada.
Flood Risk Prediction
A forward-looking Flood Risk Prediction tool is
scheduled to debut next year. It will use artificial
intelligence models to generate current and future
flood risk maps for three climate change scenarios
at the city, district, and river basin levels. The
climate scenarios are defined by the Intergovernmental
Panel on Climate Change.
The models will be trained using the inundation maps
generated by the Flood Mapping Tool and open datasets
including land use, land cover, precipitation,
temperature, gender, and age-disaggregated socio-economic data.
Together, these tools will improve the coverage of national
and regional flood early warning and risk management systems.
The system will also help build the capacity of flood
forecasting centres in the Global South to use artificial
intelligence models, big data and cloud computing to
analyse the impacts of climate change. This will be done
through hands-on training conducted at various water
and natural disaster-related conferences, webinars,
and summits. The UN University’s Canadian-based Institute
for Water Environment Health online courses.
Provided by the IKCEST Disaster Risk Reduction Knowledge Service System