►
From YouTube: 2020 CESM Tutorial - Ocean Modeling Q&A
Description
Day 2 of the 2020 CESM Tutorial featuring Ocean Modeling Questions and Answers.
The CESM Tutorial will consist of:
Lectures on simulating the climate system, practical sessions on running CESM, modifying components, and analyzing data.
For more information:
http://www.cesm.ucar.edu/events/tutorials/2020
Community Earth System Model (CESM) is a fully-coupled, community, global climate model that provides state-of-the-art computer simulations of the Earth's past, present, and future climate states.
A
B
B
B
C
D
D
D
C
Yeah,
no,
I
like
it.
When
you
get
a
chance,
can
you
send
me
the
template
for
it
and
because
I
I
know
that
we
have
to
like
start
thinking
about
our
evaluate
you
know
like
next
goals,
and
so
I
want
to
be
able
to
offer
these
kind
of
standardized
things
for
other
meetings
that
we're
doing
online
across
the
lab.
D
C
D
C
C
C
D
C
C
B
D
D
B
B
C
D
B
C
Slides
I'm
going
to
send
you
a
message
that
chloe
is
still
having
problems
so
I'll,
send
you
her
email
to
see
how
you
might
help
her
and
then
the
other
question.
You
know
we
had
that
staff.
Google,
google
chat
yesterday.
Do
you
want
to
be
in
on
that?
Even
though
you're
not
an
official
office
hours
person
today,.
E
B
B
D
D
C
C
D
C
B
E
E
E
B
B
A
C
B
So
no
so
it's
like
right
now
in
this
particular
link,
like
you
can
still.
Apparently
yesterday
you
could
still
upvote
a
downvoted
question
and
so,
as
we're
uploading,
a
down
voting
question
that
kept
moving
while
cecile
was
trying
to
ask
the
panelist
so
like
what
what
what
happened
and
so
like,
like.
I
just
asked
this
question
and.
D
A
A
C
A
C
F
A
F
B
B
B
B
A
B
B
What
less
I
pretend
to
be?
Okay,
it
is
9.
00
am
good
morning,
everybody,
I
hope
you're
all
doing
well,
and
you
had
a
fun
day
yesterday
and
I
realized
I
have
a
few
announcements
to
make
before
we
start
with
our
first
session
of
the
q,
a
power
session
regarding
the
ocean
and
biogeochemistry
and
there's
a
slight
change
in
the
agenda
just
to
allow
for
this
type
of
announcement
to
take
place
in
the
morning.
B
regarding
the
q,
a
session
to
try
to
prevent
what
happened
yesterday
when
cecil-
and
I
was
asking
questions
and
have
them
go
up
and
down
all
the
time
we're
going
to
try
to
freeze
the
questions
half
an
hour
prior
to
the
beginning
of
the
session,
so
try
to
go
through
the
question
beforehand
and
upvote
or
download
them
before
the
session
starts
about.
So
we
received
a
bunch
of
very
constructive
feedbacks
and
thank
you
guys
so
much
for
doing
this,
and
so,
as
I
mentioned
yesterday,
we're
learning
and
adapting
as
we
go.
B
It's
our
first
timer
for
the
lunch
networking
break
that
we
tried
to
set
up
for
you
for
you
to
have
time
to
meet
among
each
other.
We
didn't
plan
any
activities
per
se
so
that
you
could
use
this
time
for
you
and
we
were
kind
of
hoping
that
you
would
self-organize
and
one
way
to
do
this
is
try
to
maybe
nominate
a
moderator
for
the
session.
B
We're
gonna
try
to
have
this
so
try
to
prepare
two,
two
slight
presentation:
one
about
your
work
and
one
about
yourself
and
think
about
this.
As
an
elevator
speech,
it
shouldn't
be
highly
detailed,
very
high
level.
Think
about
that.
Your
audience
doesn't
know
anything
about
what
you're
working
on
so
avoid
the
details
and
we're
gonna
have
to
try
we're
gonna.
B
Have
we're
going
to
try
to
organize
a
speed
dating
presentation
and
see
how
it
works
and
we
will
lay
out
the
ground
rules
tomorrow
regarding
office
hours
and
I'm
very
sorry
about
the
confusion
in
and
what
happened
yesterday.
We,
I
did
update
some
of
the
slides
last
week
and
I
didn't
realize
that
I
forgot
that
you
guys
might
have
downloaded
the
slides
way
before
that.
So
please
try
to
download
the
latest
version
of
the
practicals
and
we
should
have
the
updated
common
lines
and
project
code
and
queue
specification.
B
Also,
if
you
can
try
to
go
through
a
practical
session
before
the
office
hours
start
and
you
starting
we're
going
to
send
out
an
email
today
that
will
give
you
links
to
ask
questions
for
office
hours
beforehand.
Similarly,
to
the
what
we
do
for
the
q
a
session
right
now
and
this
way,
maybe
if
there
are
general
questions
about
the
practicals
that
that
are
not
necessarily
meant
for
you
to
move
forward
with
the
practical,
then
we
could
have
someone
answer
them
in
small
breaker
rooms.
B
During
the
office
hours
section,
you
will
receive
those
links
at
some
point
today.
That's
that's
all
the
announcements
I
have
to
make.
So
I'm
going
to
stop
sharing
my
screen
and
we're
going
to
go
ahead
and
now
start
with
the
q
a
panel
session
of
the
day,
and
it's
going
to
be
moderated
by
gustavo
marcus,
who
is
a
project
scientist
in
the
oceanography
section
and
also
the
science
liaison
for
the
ocean
model
working
group.
B
His
research
is
focused
on
understanding
physical
processes,
controlling
the
ocean
circulation
around
antarctica
and
he's
particularly
interested
in
the
role
that
these
processes
play
in
regulating
the
melting
of
ice
shelves
and
the
formation
and
export
of
bottom
water.
So,
with
this
I'll
yield
the
floor
to
gustavo
and
enjoy
this
session.
A
A
A
She
is
a
postdoctoral
fellow
in
the
ocean
section
working
with
biogeochemistry
and
we
also
have
keith
lindsey,
which
is
an
expert
in
bgc,
and
you
can
also
ask
about
any
other
components
that
keith
will
know
the
answer.
So
thank
you
all
for
participating
and
I
think
we
can
go
ahead
and
start
jumping
right
on
the.
A
A
A
F
F
So
in
that
situation,
if
you
get
a
part
a
bit
of
water
from
the
upper
warm
salty
water
that
goes
down
the
warmth,
there
can
diffusion
diffuse
away
very
quickly,
so
the
temp
the
temperature
goes
down.
So
the
water
goes
from
warm
and
salty
to
cold
and
salty.
So
it's
denser
so
then
it
can
continue
to
go
fall
down
in
the
water
column.
F
This
is
producing
excess
salt
down
below
and
produces
what's
called
salt
fingers,
and
this
can
be
reproduced
in
the
laboratory
quite
easily,
and
in
that
document
I
told
you
about
this.
A
nice
picture
of
salt
fingers
happening
now.
The
second
case
is
the
diffusion
case,
and
in
this
case
we
have
cold
fresh
water
over
warm
salty.
F
So
in
this
time
case,
the
salt
distribution
is
stable,
with
fresher
water
over
salty
water,
but
the
temperature
gradient
is
unstable,
with
cold
over
warm,
but
the
profile
is
still
unstable.
I
mean
it's
still
stable,
so
if
some
water
this
time
goes
down,
remember
it's
cold
fresh.
So
it
goes
down
so
by
the
quick
diffusion
of
heat.
It
can
warm
up
there.
F
F
There's
a
equation:
204
214
in
the
pop
reference
manual
and
the
precise
formulae
for
how
that
parameterization
is
done,
are
up
there
and
it's
in
terms
of
a
quantity
called
r
sub
rho,
which
is
the
ratio
of
alpha
delta
t
over
beta
delta
s.
So
it's
the
ratio
of
which
of
the
temperature
or
salinity
vertical
radians
are
dominating
to
keep
the
ocean
stable.
A
A
G
Well,
the
the
choice
of
virtu.
I
think
there
was
another
question
further
down
a
similar
question.
The
choice
of
virtual
soft
flux
actually
is
a
bit
of
a
legacy
issue
that
earlier
versions
of
the
brian
cox
sumner
ocean
models,
which
of
which
pop
is
a
descendant
used.
A
rigid
lid,
surface
boundary
condition.
So
the
sea
level
couldn't
change.
G
G
For
a
couple
of
reasons,
one
is
that
the
the
pop
solution
of
the
barotropic
mode
uses
a
linearization
of
the
surface
boundary
condition
on
on
mass,
and
that
requires
that
the
surface
grid
cell,
the
thickness
of
the
surface
grid
cell,
be
large
compared
to
the
displacements
of
the
free
surface,
and
so
that
becomes
problematic.
If
you're,
adding
and
removing
mass
from
the
free
surface,
you
can
get
into
situations
where
the
displacement
from
the
forcing
would
exceed
the
numerically
allowable
displacement
of
the
the
free
surface.
G
So
we've
stuck
with
the
virtual
soft
flux
in
mom
6.
We
are
moving
to
a
natural
boundary
condition.
So
the
surface
is
a
material
surface
with
respect
to
salt
and
we
exchange
fresh
water
as
to
how
you
can
represent
sea
level
change
in
the
ice
age
sort
of
scenarios.
We
can
only
do
that
indirectly,
and
we
can't
do
it
in
a
continuous
integration
by
removing
water
from
the
ocean.
G
Re
re
redo,
your
topography
expose
some
more
continental
shelves,
reinitialize
the
ocean
and
go
forward,
and
I
think
that's
the
way
the
paleo
group
has
done
this
in
the
past
is
sort
of
these
time.
Slice
representations
of
the
evolution
of
sea
level
through
the
tens
of
thousands
of
years
over
a
glacial
cycle.
G
If
you
want
a
detailed
derivation
of
the
differences
between
the
natural
boundary
condition
and
salt
flux,
boundary
condition,
there's
an
appendix
in
a
paper.
First
author
is
you,
hang
sang
t-s-e-n-g
in
ocean
modeling
in
2016
about
the
representation
of
river
runoff
in
the
ocean
model
and
there's
a
careful
derivation
of
the
approximations
used
in
the
virtual
soft
flux,
boundary
condition
in
that
paper.
F
G
A
A
A
A
A
Okay,
yes
great
so
first
there
are
many
flavors
of
the
ale
framework
and,
like
broadly,
a
refers
to
any
method
that
considers
the
moving
of
the
cell
boundaries
and
in
particular,
the
way
we
we
do
imam.
6
is
just
in
the
vertical
and
this
diagram
kind
of
summarizes
what
the
a
framework
does.
So,
basically,
you
predetermined,
a
vertical
grid
that
you
would
like
to
use
to
represent
your
your
ocean
in
the
vertical.
So
in
this
cartoon
are
the
the
black
lines.
A
A
C
A
A
lagrangian
step,
and
so
every
ocean
model
that
does
that
can
be
called
a
has
a
framework.
Now.
What
is
different
about
the
one
in
bomb
six
is
that
it
includes
an
additional
regretting,
remap
step,
in
which
you
tell
them
all,
though,
that
every
certain
amount
of
time,
let's
say
a
day,
you
go
from
these
distorted
vertical
started
lines:
the
black
lines
back
to
the
predefined
line,
which
you
did
it
here.
So
you
what
you
do
is,
then
you
use
high
order.
F
F
If
you
have
convergence
of
water
into
that
volume
from
the
sides,
then
that
water
has
to
go
out
either
at
the
top
or
the
bottom
of
the
grid,
in
a
isopignal
or
in
an
ale
coordinate
instead
of
the
water
going
in
producing
vertical
velocities
across
the
boundaries.
It
actually
then
just
changes.
The
vertical.
C
F
A
G
Okay,
well,
I
think
in
general,
the
predictability
on
these
longer
time
scales
exist
because
there
are
processes
and
dynamics
in
the
ocean
which
evolve
on
slower
time
scales
than
synoptic
weather
time
scales
in
the
atmosphere.
So,
on
the
longer
end
of
those
time
scales,
they
are
basically
all
associated
with
the
generation
and
maintenance
of
heat
content
anomalies.
G
So
that's
sort
of
a
atlantic
multi-degatel
mode
kind
of
variability,
which
is,
I
think,
the
decadal
mode.
That's
been
found
to
provide
the
most
predictability
on
these
decadal
time
scales
on
the
seasonal
to
inner
annual
time
scales.
I
think
this
is
not
my
area,
but
my
understanding
is
that
most
of
that
predictability
is
basically
coming
out
of
the
equatorial
pacific,
thermocline
heat
content,
anomalies
and
the
variability
associated
with
el
nino
and
the
resultant
changes
in
sea
surface
temperatures.
F
A
F
A
Thank
you
frank.
Why
the
raspberry
juice
is
a
measure
of
the
area,
resolving
scales
I
can
try
addressing
and
then
someone
can
continue.
So
the
way
I
think
about
this
is,
if
you
have,
you
have
a
perturbation.
So
let's
say
you
have
a
instability,
a
bar
click,
instability
and
jenny.
It
generates
an
eddie
and
then
that
adi
is
being
balanced
between
buoyancy
and
and
rotation.
So
the
raspberry
ranges
is
essentially
a
a
ratio
between
who
you
see,
effects
and
rotational
effects
and
then
the
first
raspberries
of
deformation.
A
G
Well,
I'd
say
I
mean
the
the
the
dominant
scale
of
eddies
in
the
ocean
is
near
the
rossby
radius,
because
that
is
the
scale
at
which
the
instability
grows.
Most
rapidly,
the
you
know
sort
of
the
simple
edie
baroclinic
instability
problem,
the
classic
80
problem.
The
maximum
growth
rate
is
something
like
1.6
times
the
raspy
radius,
so
those
are
the
perturbations
that
are
growing
most
rapidly.
So
those
are
the
ones
that
you
see
most
readily.
F
So
the
strength
is
that
that's
the
key
is
the
barricade.
Instability
fastest
growing
is
like
a
few
times.
The
rossby
radius
and
the
raspberry
radius
is
far
larger
in
the
atmosphere
in
the
ocean
because
the
in
the
atmosphere,
the
density,
goes
from
the
surface
value
to
almost
zero
high
up.
So
comparatively
it
has
a
much
stronger
vertical
density
gradient
than
the
ocean.
A
A
E
E
There
are
other.
There
are
options
within
cesm
to
control
the
coupling
sequencing.
One
of
them
is
referred
to
as
tight
coupling,
and
I
think
it
changes
the
answers
to
this
question,
but
I'm
not
familiar
with
the
details
of
how
tight
coupling
works.
I
do
know
that
there
are
some
options
that
modify
how
the
lags
occur
between
the
ocean
and
atmosphere.
A
F
G
A
slab
ocean
or
a
mixed
layer.
Ocean
model
do
not
represent
the
horizontal
transport
processes
in
the
ocean,
so
they
can't
represent
ocean
heat
transport
directly.
You
may
with
it
with
those
classes
of
models,
you
might
prescribe
a
meridianal
heat
transport.
That's
been
diagnosed
from
another
experiment,
with
a
full
ocean
model
or
observations,
but
those
models
don't
include
the
dynamics
to
represent
those
processes.
G
E
E
Works,
I
think
there
are
runs
that
use.
You
know.
Another
option
is
to
where
so
there's
an
active
ocean
model
that
has
full
dynamics
and
that's
what
us
in
the
oceanography
section
mostly
run
so
and
then
there's
this.
The
alternative
is
a
data
ocean
model
in
one
in
there
are
comp
sets
that
use
the
data
ocean
model.
E
E
E
E
E
E
A
A
C
I
just
wanted
to
bring
up
on
this
one.
I
saw
there's
a
paper
by
shannon
wang
that
where
she
does
something
very
similar
to
this
running
csm
with
a
phytoplankton
functional
type
representative
of
phaocystis,
which
which
releases
dms
into
the
atmosphere
and
then
has
a
feedback
on
clouds,
and
so
it
has
a
cooling
effect
and
so
yeah.
This
is
published
that
somebody
has
done
something
like
this.
E
E
And
I'm
not
sure
if
they've
transitioned
that
ported
the
dms
capability
into
the
marble
framework
that
we
are
now
using
and
so
we've
sort
of
have
there
had
been
plans
to
incorporate
this,
but
that
has
sort
of
that's
not
one
of
our
top
priorities.
Right
now,
I
imagine.
If
someone
was
interested
in
pursuing
this,
then
we
could
get
the
proper
people
together
to
facilitate
getting
the
prognostic
dms
the
sulfur
capability
into
marble,
and
then
the
coupling
could
be
done,
as
was
done
by
the
lionel
doe
group.
A
Thank
you,
christian
key
next
question:
ecs,
I'm
able
to
adopt
different
time
steps
over
different
regions,
for
example,
if
we
do
refinement
in
one
region,
which
requires
a
small
time
station,
the
instability
can
use
a
large
time
step
for
all
the
regions.
With
coercive
resolution,
the
answer
is
no
not
in
the
ocean
model,
so
both
of
osha
models
that
we
have
right
now,
pop
and
mom
are
curvy.
Linear
use
the
curvy
linear
horizontal
grids.
So
there
is
some
small
refinement
as
you.
You
move
in
different
regions
and
not
much
not
like
an
unstructured
grid.
A
What
is
the
spin
up
time
for
the
ocean?
Peter
already
addressed
that
question,
so
basically,
you
can
use
the
vertical
diffusivity,
which
is
what
is
dictating
how
the
ocean
interior
is
mixing
and
divide
that
by
the
or
actually
you
you
divide.
This
is
the
square
of
the
the
depth
divided
by
the
vertical
diffusivity
in
the
interior.
You
give
your
approximate
time
scale
for
the
ocean
interior
equilibration,
remembering
that
you
use
a
a
dynamical
vertical
diffusivity
rather
than
a
molecular
one.
A
A
C
G
Clarify
that
we're
not
explicitly
representing
the
tides,
we
don't
have
a
tie
generating
force
in
pop.
I
think
there
is,
or
there
are
plans
to
put
a
tie
generating
force
in
mom
6
so
that
it
can
explicitly
represent
the
barotropic
tide
and,
to
some
extent,
the
transformation
of
the
barotropic
tied
into
internal
tides.
A
A
Mom
6
is
a
ocean
general
circulation
model
that
has
been
developed
for
climate
purposes,
so
things
that
peter
were
was
mentioning
earlier
that
you,
you
want
to
have
small
numerical
mixing
in
the
ocean
interior,
so
mum
6
has
been
designed
with
that
in
mind,
so
it
is
by
default,
designed
for
these
long
integrations
periods
in
which
you,
your
numerics,
are
good
enough
to
do
not
affect
your
solution
too
bad.
So
that's
an
advantage
for
climate
studies
compared
to,
for
example,
roms
another
one.
A
Is
you
want
to
have
a
consistency
between
your
large
scale
model
with
your
original
scale
model,
so
having
the
same?
For
both,
it
is
an
advantage.
In
my
opinion,
we
do
have
the
ao
framework
emom6,
as
we
talked
earlier,
and
that
is
another
advantage
and
finally,
more
from
a
development
point
of
view.
Mom
6
is
a
much
easier
and
faster
pace.
The
model
to
develop
because
mom
6
does
not
have
something
like
an
a
joint
model
like
realms
in
which
makes
it
much
harder
to
develop
the
model.
A
Every
time
you
need
to
change
something
you
also
need
to
change
the
adjoin
for
that,
which
is
a
complicated
piece
of
code
in
terms
of
design
advantages.
What
I
can
think
of
is
mom.
6
is
not
as
widely
used
for
the
coastal
physical
oceanography
community
as
realms,
and
also
the
open
boundary
conditions,
which
is
a
very
important
piece
within
the
regional
framework.
A
It's
still
underdeveloped
the
development
in
mount
six,
as
we
speak
at
the
latest
meeting
that
people
are
doing
that
kind
of
work,
the
results
are
pretty
good
between
moms
and
mom
and
rome's,
but
it
is
not
the
open
bonding
conditions
imam
is
not
as
well
established
as
it
is
in
roms.
I
would
say
right
now
any
other
points.
A
E
There
is
also
a
air
c
flux
of
oxygen
with
the
atmosphere
based
upon
the
saturation
solubility
of
oxygen
and
seawater,
which
is
a
temperature
and
salinity
dependent
computation
so
that
cold
water
holds
more
oxygen
and
if
the
water
is
under
saturated,
then,
with
respect
to
the
equilibrium
saturation
concentration,
then
the
ocean
takes
up
oxygen
from
the
atmosphere
and
vice
versa.
If
the
ocean
is
super
saturated
with
respect
to
that
equilibrium,
saturation
concentration.
E
The
atmospheric
boundary
condition
is
fixed
in
oxygen
of
the
amount
of
oxygen
that
goes
in
and
out
across.
The
erg
exchange
is
very
small
compared
to
atmospheric
concentrations
and
we
treat
the
atmospheric
concentration
of
oxygen
as
as
a
fixed
value.
So
we
don't
change
the
atmospheric
concentration
of
oxygen
when
we
have
air
sea
exchange.
C
A
B
F
Well,
gustavo,
I
think
you've
missed
out
the
one
about
the
the
densities.
Thank
you.
So
the
answer
to
that
is
that
we
use
the
most
accurate
one
which
is
c
described
in
the
in
the
manual
and
the
most
accurate
one
has
been
speeded
up
quite
a
lot
recently
well
recently
about
10
years
ago.
So
no
reason
not
to
ten
years
ago
is
recent
in
my
calculation.
F
G
Well
I'll
just
say
that
we've
removed
it
from
the
ocean
in
the
most
recent
versions
of
cesm,
because
it's
not
connected
to
the
rest
of
the
global
ocean.
It's
a
basically
a
lake
and
for
many,
in
fact
the
I
think
it's
may
still
be
the
case.
The
topography
is
completely
wrong
in
cesm,
because
the
national
geophysical
data
center's
topography
database
had
the
level
of
the
surface
of
the
caspian
sea
for
its
depths.
G
E
So
for
the
tracers,
the
passive
tracers,
this
is
a
work
of
this
is
active
research
and
there
was
actually
a
webinar
that
I
gave
on
this
topic.
Yesterday.
Elizabeth
included
a
link
to
the
webinars
in
the
chat.
I
don't
think
my
talk
has
been
posted
there
yet,
but
their
software
under
development
for
spinning
up
passive
tracers.
E
E
E
You
get
more
years
per
day
in
an
ocean
ice
configuration
using
forcing
from
a
coupled
model
than
you
can
when
you
run
with
the
fully
coupled
model.
So
I'm
not
certain
how
this
question,
if
they
were
thinking
about
spinning
up
tracers
to
equilibrium
or
just
getting
more
years
per
day
in
a
particular
configuration.
So
there
are
work
in
progress
on
the
former
and
you
can
do
the
ladder
of
getting
forcing
from
the
coupled
model
and
running
ocean
ice
configurations
with
that.
Forcing.
A
C
A
C
A
Yeah,
I
think
that's
a
question
for
yeah
the
polar
q,
a
they
will
be
able
to
show
that
is
better
than
than
us
all
right.
So
I'm
going
to
skip
that
one.
Let's
end
the
session
with
one
more
last
question:
let's
get
this
last
one
here.
What
is
the
method
that
seam
coupler
uses
to
couple
the
ocean
and
atmosphere?