Delta Life Nr. 4 September 2015

26 / 36

26
Twitter
Modern forecasting systems combine up-to-date
data fromweather stations and information about
river discharges, groundwater and sea level with
hydrological models, establishing a basis for
predicting high-water levels in rivers, sea floods and
imminent drought.
Systems like Delft-FEWS make forecasts faster, and the fore­
casts look further ahead and are more accurate than ever. As
soon as a clear picture emerges of the risks for people, infra­
structure or agriculture, the system issues a precise warning.
This is usually a few days in advance for floods, and even a
few weeks or months for droughts. The emergence of big data
also makes it possible to draw on more and more alternative
sources of data such as Twitter and satellite images to make
the warnings even more accurate.
The result: more and more time for local authorities to deploy
the emergency services, to organise evacuations or to install
additional protection on rivers or coasts, or to create extra
water stocks to cope with a drought.
Intense rainfall
A good example of a warning system for river floods is the
National Flood Forecasting System for Great Britain and
Wales. The NFFS produces forecasts for floods from the main
rivers, reservoirs, estuaries and seas, taking surface water
and groundwater into account. It does this by combining
hydrological measurements and models with meteorological
forecasts. On that basis, it is possible to state three days
in advance whether the water-level limit will be exceeded
at important locations and what the impact will be on the
surrounding area and the local people.
Even so, the major floods in central England and Wales in
the summer of 2007 resulted in thirteen fatal casualties and
caused material damage amounting to £3.2 billion. During
the evaluation of the disaster, it was found that hydrologists
and meteorologists should have collaborated better. The
result was the Flood Forecasting Centre, a joint agency
bringing together the Environment Agency and the Met Office.
The first major test for the system was in the winter of 2014.
The authorities were able to warn local people in time and
provide them with adequate information, and so injuries and
damage were reduced substantially.
Sea-level rise
On Mauritius, an island republic in the Indian Ocean, sea-
level rise and the increasing intensity of tropical storms are
a threat: they could inflict serious damage in the future on
infrastructure, farmland and housing. Most Mauritians live
on the coastline and so they are particularly vulnerable.
To mitigate the risks of storm surges, the government of
Mauritius set up the Climate Change Adaptation Programme
in the Coastal Zone of Mauritius.
An important part of this programme is a storm-surge
warning system that will be operational for the cyclone
season in January and February of next year. The system
combines oceanography models with meteorological
forecasts. The coastal inhabitants on the side
of the islands where a storm surge is expected
will be warned three days in advance. That gives
them time to evacuate to safe places on the island
and allows the authorities to deploy their limited
resources and emergency services in the best possible
way. The prevention of fatal casualties will be the priority
here and economic damage can also be kept to a minimum.
Peat fires
Both systems operate on Delft-FEWS, a software platform
that will be celebrating its tenth anniversary this year.
Enormous progress has been made during that time. The
available data flows and hydrological models are being used
ever more efficiently. The increasing availability of satellite
data and databases with soil information has resulted in a
major boost for the warning and forecasting systems.
Delft-FEWS is a platform for processing and combining
data flows and it was initially developed for hydrological
forecasting and warning systems. A set of modules make
it possible to develop systems that are tailored to the
specific requirements of organisations in specific regions.
The system can be used in areas such as water-quality
forecasting, groundwater and reservoir management,
operational management and the optimisation of sewage
systems, and even for predicting peat fires.
Social media
A recent development involves using social media for
warning systems. The Floodtags company produces, in
collaboration with Deltares, detailed real-time flood maps of
cities in South-East Asia on the basis of tweets. The maps
show which streets are flooded when the river bursts its
banks and that in turn helps to predict the further course of
the flood.
The recently opened iD-Lab in Delft is another important
new facility where all the available data flows and computer
models can be brought together in a warning and fore­
casting system. During emergencies, the iD-Lab is set up
as a crisis centre that supports government agencies and
emergency services by supplying up-to-date forecasts and
warnings.
No end in sight
The development of forecasting systems is destined to
continue for a long time yet. Increasing amounts of infor­
mation from the social media, more detailed forecasts,
more efficient software, improved hydrological models and
support for the agencies involved: all these factors means
that forecasting will become even more valuable in the years
to come.
For more information:
Delft-FEWS:
iD-Lab: