Operational European Flood Forecasting - Environmental Modelling: Finding Simplicity in Complexity - page 411

numerical information and statistical plots, which can be
achieved in a range of useful ways. These products are
reviewed within the EFAS partner network on an annual
basis and updated according to feedback provided. The
archive of EFAS results is made publicly available and
canbe viewed at http://floods.jrc.ec.europa.eu/efas-flood-
forecasts.html.
EFAS results are visualized as maps and time series as
well as hotspot points on the EFAS web interface. The aim
is to guide the forecasters with overview products (such as
the hotspot maps) of those river basins where some prob-
ability of flooding is forecast. With more detailed map
and time-series information, either summaries composed
of the different forecasts or results from the individual
forecasts, the forecasters can start analysing the situa-
tion in detail and make a decision whether or not to
issue a flood alert. A particular challenge for EFAS is to
represent multiple forecast information in a concise and
easy-to-understand way.

map with some additional information. This represen-
tation was inspired by the way meteorologists visualize
station data and is very useful as a first assessment. The
example shown in Figure 25.6 is for the EFAS forecast of 5
January 2011, 12:00. The figure indicates high probability
of widespread flooding at the border between France,
Germany and the Benelux, along the Polish-German bor-
der as well as in north-west Spain. The numbers next
to the reporting point indicate the number of ensemble
members exceeding the EFAS high threshold; thus, at
these locations there is a probability forecast that dis-
charges exceed a five-year return-period threshold. The
left number is for the global ensemble, the right number
for the COSMO-LEPS ensemble.
Similar maps are produced for the different individual
forecasts or combined information. Examples here are a
map illustrating in which river stretches both determin-
istic forecast exceed a threshold.

Time-series information

Spatial overview maps
The first map visualized when connecting to the inter-
face is a hotspot map (Figure 25.6). Following a decision
tree including criteria such as minimum upstream area,
minimum probability, persistence of signal with regard
to previous forecasts, a control point is plotted on the

At each of the hotspot points, time series can be visu-
alized. In order to be able to capture the information
from multiple forecasts in the best way, different prod-
ucts are produced. One of the classical representations
of probabilistic forecasting standard box-plot diagrams
with typical quantile intervals of minimum/maximum,

Figure 25.6 EFAS hotspot map from the
forecast of 5 January 2011, 12:00.
Forecasters can zoom into the area of
interest and explore other products, such as
spatial maps showing the probability of
exceeding the EFAS High threshold
(Figure 25.7). The probability maps
highlight the highest probabilities to exceed
the EFAS high threshold for the entire
forecasting leadtime - in this example the
highest probabilities are within the smaller
tributaries of the Mosel/Rhine basin near
the French-German border whereas in the
main rivers the probabilities are low (From
EFAS).