Operational European Flood Forecasting - Environmental Modelling: Finding Simplicity in Complexity - page 407

Table 25.2 Setup and datasets required for a European operational flood forecasting system.

Data type Data set/activity Specification

Weather forecast Meso-scale meteorological

EPS with leadtimes of up
to 15 days or more, twice a
day

Necessary to achieve medium-range flood forecasts, highest resolution
possible for best quantification of severe rainfall events, high number
of EPS necessary to capture extreme weather situations

Monthly forecasts (30 days),
at least once a week

For early indication of potential flood threats

Reforecasts, once a week Necessary to calibrate EPS and reduce systematic bias

Short-range EPS, e.g. with
different lead times.

Necessary for flash-flood-type events and to improve quality of forecasts
at the onset of a crisis or during a crisis for improved crisis
management support to the partners

High-resolution
deterministic forecasts

Spatial resolution should be higher that of the EPS and provide leadtimes
of at least 7 days

Observational Real-time meteorological Usually meteorological data of rainfall, temperature, wind cloud cover

evaporation and relative humidity amongst others are required. The
station density depends on the size of the domain and topography (e.g.
one requires more stations in mountainous terrain). This is needed for
example to derive the initial conditions or to correct forecasts.

Real-time hydrological Information about river discharge and/or water levels is needed to

validate model results and for real time correction of model outputs.

Real-time satellite
information

For data assimilation: variables such as snow cover, soil moisture or
precipitation for blending with observations to setup initial conditions
and/or update (correct) the model results.

Static maps GIS maps to setup hydrological model with spatial information on

topography, river network, land use, soil type and depth, lakes, etc

Historic meteorological and
hydrological

For the calculation of model climatologies, calibration and validation of
results, case, generation of flood warning thresholds and case studies.

Hydrological model Pre-processor for

meteorological forcing

Correcting any bias and downscaling from meteorological grid scale to
hydrological model grid scale

Modelling code Including hydrological processes routines and data transfer and storage

Post-processor for river
discharge results

May include forecast correction (updating) with observed river discharge
and processing of river discharge against critical flood thresholds

Model performance Model validation, calibration and sensitivity analysis routines. Providing

a range of performance scores for long historic (reforecast/hindcasted)
runs and post-flood event analysis in forecast mode.

Data processing Collection and storage Data needs to be collected from various sources, transferred, and

converted into a standard format before storing in database or file
storage

Quality control Historic and real time data require different levels of quality control.

Identification of outliers, erroneous reporting and reliability of station
data or other sources need to be checked

Data extraction Tasks include temporal aggregation of data to model time steps, spatial

and temporal interpolations as gridded fields in model required input
formats

( continued overleaf )