Catchment scale flow and transport: selecting model complexity and using parametrization and scaling methods to develop robust and efficient models
Presented by Professor Sabine Attinger
Department of Computational Hydrosystems, Helmholtz Centre for Environmental Research - UFZ
There
is
a
strong
need
to
develop
regional
scale
hydrological
models
that
are
predictive
under
global
change
to
help
decision
makers
in
water
resources
management.
There
are
different
ways
to
achieve
this
goal.
Often,
model
development
involves
adding
more
processes
and
using
increasingly
higher
spatial
and
temporal
resolution.
Following
this
path
it
is
assumed
that
higher
process
complexity
will
improve
the
predictability
of
models.
However,
the
complexity
of
the
model
must
be
deliberately
chosen,
with
degrees
of
freedom
appropriate
given
the
available
data.
The
more
complex
the
model,
the
more
data
are
needed
to
calibrate
the
model.
To
this
end,
another
modeling
approach
is
to
establish
less
complex
(often
called
conceptual)
or
aggregated
models,
which
have
fewer
free
parameters
and
require
less
data
for
the
model
parametrization.
Usually
these
models
suffer
from
a
lack
of
transferability
across
different
regions
and
scales.
In
my
talk
I
will
introduce
an
alternative
way
to
select
an
adequate
model
complexity
namely
from
a
goal
oriented
perspective
and
will
present
a
modelling
philosophy
together
with
a
multiscale
parametrization
method
(MPR)
that
will
result
in
models
that
are
robust
and
transferable
across
scales
and
regions.
Everyone Welcome. Coffee Provided.