►
From YouTube: CESM Workshop: Ocean Model Working Group
Description
The 26th Annual CESM Workshop will be a virtual workshop with a modified schedule on its already scheduled date. Specifically, the virtual Workshop will begin with a full-day schedule on 14 June 2021 with presentations on the state of the CESM; by the award recipients; and three invited speakers in the morning, followed by order 15-minute highlight and progress presentations from each of the CESM Working Groups (WG) in the afternoon.
On 15-17 June 2021, working groups and cross working groups have half-day sessions, some with presentations and some that are discussion only.
A
A
B
C
C
So
after
this
brief
introduction,
frank
is
going
to
give
a
little
status
update
and
then
we'll
have
three
talks
according
to
progress
got
it
then
we'll
have
three
talks
before
the
break
at
10
o'clock
local
time
here
we'll
have
a
20-minute
break
and
then
we'll
kick
off
a
session
on
waves
in
in
mom
six
and
that's
a
sec,
a
series
of
shorter
talks
which
will
end
at
about
10
11
25.
So
my
my
slide
here
is
not
quite
right.
I
think
we
have
one
more
talk
at
11
20.
C
This
short
11
25
will
then
have
25
minutes
of
open
discussion
on
waves
or
no
at
11
25
we'll
have
discussion
on
waves
and
then,
after
that,
a
brief,
open
discussion
on
whatever
we
want
to
talk
about.
It
should
be
done
by
just
a
couple
of
reminders,
so
we
are
streaming
this
live
on
youtube
and
then
it
will
be
recorded
and
available
for
anyone
to
watch
on
youtube.
C
Also,
just
a
reminder
that
the
text,
whatever
you
put
in
the
chat
is
that
is
saved,
can
be
saved,
including
private
messages,
not
just
ones
that
are
to
everyone,
but
ones
between
each
other.
Those
are
saved
as
well,
and
I
think
that
is
rules
of
the
road
so
keep
your
microphone
muted
during
the
presentations.
Please,
if
you
have
a
question
during
the
talk,
please
just
use
the
chat
and
then
we'll
have
q
a
after
each
talk.
Sorry
about
that
and
during
that
time,
there's
a
lot
of
people
here.
C
A
Yeah,
just
let
me
a
couple
of
addendums
on
the
agenda,
so
the
only
change
in
the
agenda
is
alpha.
A
Altoontis
will
give
a
very
brief
kind
of
a
status
update
of
where
we
are
right
now
with
respect
to
coupling
wave
watch
3
and
mom
6
at
the
as
the
last
of
those
short
talks
on
wind
wave
issues
and
leading
into
the
discussion
to
sort
of
set
the
set,
the
stage
of
where
we
are
so
we
can
be
talking
about
where
we
want
to
go
and
then
the
other
thing
is
that
we
do
have
the
zoom
call
for
another
30
minutes
after
12
o'clock.
A
So
we
we
don't
absolutely
have
to
finish
at
12
o'clock,
okay
in
terms
of
status
and
updates.
I
didn't
actually
prepare
another
presentation,
but
this
I
can
show
what
I
showed
on
the
monday
plenary
update.
If
people
have
want
to
see
something,
but
let
me
just
put
it
out
there:
the
the
the
process
on
monday
was
each
we
blasted
through
all
whatever
13
or
14
working
groups,
without
an
opportunity
for
people
to
ask
questions.
D
Frank,
I
have
a
question:
this
is
nikki
hi,
nikki
hi,
there's
a
lot
of
talk
about
wave
watch
three
and
representing
waves
in
mom
six,
and
I'm
wondering
if
that
will
there
are
plans
to
eventually
transition
that
to
the
marble
interface
in
the
sense
of
representing
bubbles
and
their
role
in
gas
exchange,
or
things
like
that.
Could
you
speak
to
that
a
little
bit.
A
I
think
that's
an
excellent
question
for
the
discussion
after
words,
I
hope
that
some
people
might
want
to
talk
about
the
role
of
waves,
breaking
waves
in
air
c
flux
calculations
and
I'm
not
the
person
to
really
I'm
not
really
informed
on
that
issue.
So
maybe
we
can
bring
come
back
to
that
at
11,
25.
A
Description
of
our
status,
so
we
taught
we
in
terms
of
our
configurations
that
we
have
available.
You
know
I
showed
on
monday
these
different
vertical
grids
we've
explored
and
much
of
that
material
was
actually
shown
at
our
winter
working
group
meeting.
A
Since
that
time,
a
couple
of
things
have
happened.
One
is
that
gustavo
and
keith
lindsey
working
with
folks
at
gfdl,
have
kind
of
reworked
the
implementation
of
passive
tracers.
So
we
are
now
we
have
cfcs
going
sort
of
as
our
first
test,
so
we'll
be
able
to
use
those
in
the
evaluation
of
vertical
grid
and
any
other
parameterizations
that
we're
bringing
into
the
model
over
the
next
months
and
years.
A
The
other
is
we
have
made
if
you
caught
peter
laritzen's
talk
in
the
amwg
yesterday
and
this
issue
that
I
brought
up
in
my
presentation
with
respect
to
energetics
of
the
system
and
just
to
expand
on
that.
A
little
bit,
if
you
recall
pop,
is
a
constant
volume
model,
so
we
don't
actually
exchange
fluid
mass
between
the
ocean
and
the
other
components
of
the
hydrologic
cycle.
A
So
when
it
rains
over
the
ocean,
we
don't
actually
add
water.
We
remove
salt,
it
uses
a
virtual
salt
flux,
boundary
condition,
so
we
remove
the
amount
of
salt
that
is
required
to
achieve
the
same
dilution
of
the
surface
layer
with
respect
to
salinity,
so
that
made
things
much
simpler,
well
somewhat,
simpler
in
terms
of
the
energetics,
because
that
water
moving
from
the
atmosphere
or
the
rivers
into
the
ocean
carries
some
enthalpy
and
we
didn't
have
to
account
for
that,
because
we
weren't
adding
any
mass
to
the
ocean
we
didn't
have
to
carry.
A
We
didn't
care
about
what
the
enthalpy
of
that
mass
we
weren't
adding
is
and,
as
it
turned
out,
the
atmosphere
didn't
care
about
it
either,
but
now
that
we
have
a
a
model
with
a
natural
boundary
condition,
we
have
to
account
for
that
that
mass
and
we
went
through
a
process
of
sort
of
evaluating
how
big
of
a
term
that
is-
and
it's
not
negligible.
You
know
locally,
it's
tens
up
to
100
watts
per
meter
squared
and
in
global
average.
A
A
So
again,
mariana
and
gustavo
have
implemented
the
first
cut
at
this
and
being
able
to
exchange
those
energy
fluxes
associated
with
mass
exchange,
but
as
of
today,
the
other
component
models
are
not
accounting
for
that.
So
we're
working
on
a
strategy
with
the
other
working
groups
on
how
how
to
deal
with
that,
and
it's
not
as
straightforward.
A
As
you
know,
the
the
complexities,
the
atmospheric
thermodynamics
are
quite
a
bit
bigger
it's
more
complex
than
our
system.
Additionally,
our
river
transport
model
has
no
thermodynamics.
There
is
so
there
is
no
way
to
exchange
energy
with
it,
so
we
will
need
to
design
and
implement
some
kind
of
a
global
fixer
to
account
for
that
energy
deficit.
A
So
those
are
some
of
the
things
we've
been
working
on
as
within
the
ocean
section,
with
our
resources
for.
C
That's
good
to
know
that
you're
doing
that
frank,
I
know
at
gfdl,
we've
been
from
the
ocean
side
trying
to
get
the
atmospheric
people
to
do
this
for
more
than
a
decade,
and
they
say
it's
a
really
difficult
problem
with
parameterizations
and
just
lots
of
details
that
are
difficult
to
deal
with.
C
A
Well,
I
I'm,
I
guess
our
initial
target
is
to
close
the
terms
that
are
straightforward
to
close
and
then
put
in
a
global,
some
kind
of
a
fixer
for
the
terms
that
aren't
so,
for
example,
river
runoff.
There's
just
no
way
to
do
that
and
if
from
first
principles
in
our
system.
C
C
A
C
A
C
E
F
Yeah,
I
just
wanted
to
add,
if
you
listen
to
peter's
talk
in
in
what
he's
doing
in
the
atmosphere,
he's
making
approximations
and
say
if
we're
below,
I
think
it's
something
like
.01
watts
per
square
meter.
Then
he
is
neglecting
it.
So,
whatever
we
do,
we
should
sort
of
have
a
consistent,
a
negligible
kind
of
number,
because
that's
what
he's
doing
so
we're
not
going
to
be
down
to
the
10
to
the
minus
six
or
a
round
off.
It's
going
to
be
something
like
that.
F
A
Yeah,
the
the
other
thing
I
heard
was
that
well,
I
didn't
hear
it,
it
was
actually
a
thread
in
the
chat
during
peter's
talk
was
that
apparently,
the
nor
esm
group
which
runs
cam
6
has
addressed
this
issue
in
their
version
of
camp,
so
I
think
it
would
be
worth
following
up
with
them
to
see
what
they've
done
all
right.
Let's
it's
nine
o'clock
ian
you
wanna
take
back
over.
G
A
First
up
is
zhang
lin
wren.
Her
title
is
the
impact
of
atlantic
meridian
overturning
circulation
on
marine
heat
waves
during
climate
warming,
so
zhang
lin.
You
want
to
go
ahead
and
share
your.
A
A
H
Hello:
everyone,
I'm
charlene,
a
first-year
graduate
student
from
the
department
of
earth
and
planet
science,
planetary
science
of
ucr,
and
this
time
I
would
like
to
share
with
you
my
recent
research
progress
on
market
with
and
include
in
collaboration
with
dr
liu
and
the
topic
is
the
impact
of
atlantic
martial,
overturning
circulation
on
marine
waves
during
climate
warming,
and
since
the
1980s
land-based
trade
waves
has
been
occurring
frequently
in
response
to
the
continuous
global
warming.
More
frequent
and
intense
market
wave
events
have
accompanied
the
rest
in
the
sea
service
temperature.
H
I
think
sometimes
marketwave
can
have
an
impact
on
nitrogen,
as
well
as
on
human
society,
and
in
recent
years
many
scientists
have
conducted
research
on
mars
waves
and
because
market
waves
are
anomalously,
warm
water
event,
increase
increased
sea
service
temperatures
greatly
increase
in
the
lack
of
market
waves.
The
figure
on
the
right
shows
otherwise
shows
the
result
based
on
the
ost
vertical
observations.
H
The
global
mean
surface
temperature
continues
to
increase
during
the
global
warming
process
and
correspondingly
the
frequency,
duration
and
intensity
of
market
waves
show
an
increasing
trend
and,
generally
speaking,
grooming
has
led
to
a
higher
fancy,
wider
text
range
and
greater
intensity
of
microwaves
and
next,
some
introduction
to
about
the
emog.
The
amok
is
a
major
machinism
for
hatred,
redistribution,
planet
and
an
important
factor
in
climate
variability
and
a
mob
is
a
sensitive
system
and
on
the
satellite
temperature,
salt
and
density
differences
in
the
ocean
and
the
right
panel
shows
the
change
social.
H
The
change
in
aim
of
over
time
in
the
clamped
model,
where
the
dark
blue
curve
indicates
multiple
modal
mean,
and
we
can
see
that,
after
a
long
and
relatively
stable
period,
beginning
in
1960,
amok
experienced
a
rapid
decline
in
the
mid
20th
century
and
the
weakening
of
ammo
is
also
confirmed
by
the
first
observations.
H
Will
the
slowed
ammo
have
an
impact
on
the
market
website
has
been
increasing
and
if
it
is
an
impact,
how
do
the
marine
waves
change
under
under
the
weekend
they
move
and
how
does
a
move
affect
the
marine
waves?
These
are
the
focus
for
the
focus
of
of
our
research.
H
Our
aim
is
to
explore
whether
the
emote
will
affect
the
merit
waves
and
the
global
warming.
We
need
to
control
for
the
changes
of
and
changes
in
the
emote,
and
we
use
the
fact
example,
members
from
the
season
4
historical
and
rcp
8.5
simulations
and
focusing
on
daily
sea
surface
temperature
information.
H
The
study
time
period
as
1081
to
2020
and
2061
to
2100
to
explore
the
effects
of
emote
on
marital
wave.
The
result
of
a
set
of
experiments
in
mogfix
were
chosen
as
a
control
in
this
experiment.
H
Fresh
water
removed
from
the
vast,
deep
water
information
zone
in
the
south,
polar
north
atlantic,
the
blue
area
and
as
amok
inter
intel,
remains
almost
stable
to
1980,
where
global
warming
continues,
the
data
used
used
in
this.
They
are
shown
on
the
right
and
with
right
parts
representing
the
data
for
fixed
emote
and
the
blue
part
for
normal
conditions.
A
solid
curve
indicates
multiple
modal
mean,
and
it
is
clear
that
emote
is
weakened
under
natural
conditions.
H
H
Look
at
the
completion
of
my
waves
first
to
define
the
magnetic
waves.
A
market
wave
is
defined
as
a
discrete
long-term
anomalously
warm
water
event
that
occurs
at
a
specific
location
to
explore
some
of
the
characteristics
of
market
wave
scientists
quantify
this
quantifies
general
definition
and
first,
it
is
based
on
whether
the
sea
surface
temperature
exceeds
a
set
threshold
to
determine
whether
it
is
an
abnormally
warm
event
and
to
determine
the
threshold.
The
clientology
average
is
first
calculated.
H
H
Since
subsequent
studies
performing
a
merit,
wave
threshold
are
image
like
the
panel
b,
which
use
daily
temperature
values
for
all
years,
with
an
11d
window,
centered
on
a
particular
day
and
calculates
a
19
percentile
threshold
for
each
day
and
then
performs
a
31
days.
Smoothing
and
this
method
is
ensures
a
sufficient
sample
size
and
the
threshold
will
vary
seasonally
after
we
identify
market.
We
based
on
the
delays
of
temperature
time
series
and
we
can
characterize
their
frequency,
intensity
and
duration
duration
is
defined
as
the
time
between
the
start
and
end
date
of
the
event.
H
H
Past
40
years
and
the
left
and
the
right
panel
is
the
result
of
the
model
over
in
the
observations
and
the
ocean
model
are
similar
and
in
general,
the
model
can
be
used
to
simulate
the
characteristics
of
marine
groups
in
the
real
ocean
and
in
order
to.
H
Investigate
the
global
distribution
of
market
waves
and
its
development
under
different
intensity
aim
of
effects
210,
where
period
are
selected
for
an
analysis
and
the
first
one
is
1981
to
2020
as
show
in
the
figure.
The
first
column
is
the
result
of
natural
case.
The
immo
cream
and
second
volume
is
a
result
of
the
immobilic
experiment.
The
last
column
is
differences
between
the
two
previous
ones.
It
will
fix
miners
not
immigrate
during
the
last
40
years.
H
Then
appear
mainly
in
the
western
boundary
currency
mission
and
central
undertake,
with
an
average
annual
frequency
of
about
three
or
two
to
three
times
and
a
duration
about
twice
as
long
as
under
airborne
weakening
conditions
are
observed
and
the
effects
of
aim
of
changes
are
merely
intensity
is
not
significant.
Significant
in
this
time
period
and
only
weak,
only
enhancements
are
observed
in
a
few
coastal
regions
and
the
above
result
indicator
sides.
H
H
South
of
the
green
line,
this
area
and
stability
on
the
duration
and
frequency
of
mercury
wave
is
opposite
in
the
northern
and
southern
hemisphere.
In
most
of
the
northern
hemisphere.
Stronger
emote
is
accompanied
by
longer
marine
heat,
wave
duration
and
increased
frequency,
while
shorter
duration,
market
width
and
lower
neural
mean
frequency
of
market
wave
will
occur
in
the
southern
hemisphere.
H
Frequency
of
the
result
of
the
first
40
years
of
1981
to
2020
and
between
amokfix
and
mc3,
the
variation
of
a
mold
is
much
stronger
and
the
variation
of
our
heat,
wave,
duration,
intensity
and
a
neural
mean
frequency
is
also
much
stronger.
H
H
North
atlantic,
over
time
for
two
period
is
displayed
for
the
first
attempt
in
1981-20.
The
result
of
the
two
data
sites
are
closed
and
the
result
of
input
fix
are
slightly
higher
than
the
ammo
free
in
the
last
last
last
decades,
and
during
2061
to
2000
american
twists
and
their
ammo
fixed
conditions
are
much
stronger
than
those
under
immunog3,
with
overall
increasing
trend
in
generation.
H
H
The
extremely
long
duration
of
market
wave
at
this
location
leads
to
a
lower
calculation
of
the
frequency,
and
this
slightly
decrease
is
observed.
After
the
overaverage,
the
global
average
compare
the
left
and
right
column
plots.
It
is
one
side.
The
gap
between
the
two
right
and
blue
curves
on
the
right
side
is
larger.
H
This
is
this
may
be
due
to
the
fact
that
the
emote
has
been
continually
weakening
since
1980
and
the
right
hand.
Side
right
hand
corresponds
to
a
small,
smaller
flow
of
evil
than
the
left
hand.
So
why
does
the
strings
of
a
mob
affect
marine
heat
wave?
And
we
guess
that
it
may
be
related
to
the
effect
of
emote
on
sea
surface
temperatures.
H
If
the
whole
distribution
changes
towards
warmer
climate
and
there
will
be
more
warm
events
and
less
cold
events,
and
if
the
temperature
variability
increases
without
mean
change,
there
will
be
more
warm
and
cold
events
and,
in
addition,
if
the
distribution
pattern
changes,
there
will
be
as
astronautic
as
astronautic
changes
of
warm
and
cold
event.
The
three
regions
we
selected
are
the
in
this
area:
the
mediterranean
and
the
northwest
atlantic,
where
marine
feed
with
have
have
occurred
before
and
the
region
south
of
greenland,
which
is
the
north
atlantic
warming.
H
Hope
in
this
area.
Overall,
the
sisters,
creature,
anomaly
probability,
density
function
of
a
mog
fix
in
all
three
regions
have
physically
paneled
out
in
the
wood
direction
compared
to
the
input
three
and
the
distribution
has
changed,
with
a
significant
increase
in
extremely
warm
events.
The
degree
of
action
is.
H
Much
in
the
later
40
years,
81
and
2020
the
mean
series
temperature,
anomaly
values
in
the
three
region:
increased
by
0.05
to
0.1
degrees
celsius
and
in
the
period
26
1
to
2100
the
mean
surface
temperature
normally
increases
by
0.5
to
3.8
degrees
celsius.
H
It's
larger
than
the
first
time
period
and
in
addition,
it
can
be
seen
from
the
panel
seaside.
The
left
and
right
side
of
the
invoke
fixed
curve
are
not
symmetrical
in
the
region
south
of
finland
with
more
in
the
right
hand
and
which
indicates
a
higher
increase
in
the
warming
events
and
in
general,
the
increase
in
mean
surface
temperature
is
one
of
the
main
reasons
for
increase
in
male
heat
waves
and
the
higher
latitude
region
have
a
higher
mean
surface
temperature
increase
and
hair
system.
A
A
No
well
that's
unfortunate.
We
are
at
the
end
of
the
time
I
see
that
her
advisor
is
on.
Maybe
you
would
be
willing
to
entertain
a
question
or
two
based
on
the
material
we
we
did
get
through.
I
Sure
sure,
thanks,
yes,
okay,
I
think
charlie,
has
has
been
got
through
all
the
most
materials
and
basically
only
like
the
final
slides
for
the
conclusion
so
yeah.
Thank
you
very
much.
I'd
like
to
to
the
question.
I
saw
the
chat
box.
I
How
does
the
front
shell?
So
how
does
the
inbox
fix?
So?
Basically,
we
will
fix
emog
by
extracting
fresh
water
from
the
south
polar
atlantic
covering
the
deepwater
formation
area.
So
by
doing
that,
we
kind
of
like
to
control
the
water
formation.
Basically,
we
can
maintain
the
immoral
strength
under
global
warming
scenario-
rcp
8.
8.5
for
this
one.
So
that's
basically
what
we
did
so
we
published
paper
in
2020
last
year,
so
basically
more
details
can
be
from
that.
I
Another
question
from
nikki
is
the
fixed
mortgage
symbol
output
available
yeah,
so
we
have
just
tried
to
pull
the
data
and
so
yeah
if
you're
interested.
Please
contact
me
so
high
resolution
yeah
that
that
that'd
be
great.
If
we
can
see
the
higher
resolution
definitely
and
yeah,
and
also
we
realized
the
high
resolution
model.
So
we
compared
with
low
resolution
and
it
could
be
greatly
affected
by
the
resolution,
especially
in
the
coastal
area,
so
that'd
be
that'd,
be
very
important.
Thanks.
C
J
C
Do
you
want
to
go
ahead
and
share
and
patrick
is
going
to
talk
to
us
about
evaluating
coupled
climate
model
parametrizations
via
skill
at
reproducing
the
monsoon
intra-seasonal
oscillation
thanks.
K
I
think
my
sound's
on
baylor
taught
me
that
you
can
change
the
title
of
the
talk
if
no
one
notices
so
there's
that
hi
again,
I'm
patrick
ornstein
and
I'll
be
presenting
on
evaluating
turbulence
parameterizations
by
their
impact
on
the
model.
Monsoon
intra-seasonal
oscillation.
K
Just
to
outline
I'm
going
to
describe
the
model
parameterizations,
we
investigated
briefly
described
the
monsoon
intra-seasonal
oscillation,
just
to
give
quick
context
and
present.
Some
of
the
results
of
our
model
runs
and
then
discuss
some
of
the
implications.
K
K
And
they
tend
to
show
the
mixed
layer.
The
figure
on
the
bottom
is
from
fox
kemper
and
collaborators
2008,
and
it
shows
a
temperature
front
and
an
idealized
simulation
of
the
upper
ocean.
The
black
arrows
are
indicating
where
the
mixed
layer,
eddies
seen
at
the
surface,
are
causing
the
front
to
overturn
and
re-stratify.
K
The
miso
is
a
mode
of
variability
of
rainfall
in
south
asia
which
acts
on
a
shorter
time
scale
than
the
indian
monsoon
itself
to
it's
best
illustrated
if
you
look
at
the
figure
on
the
right,
which
is
from
a
2016
oceanography
paper
and
shows
three
representative
monsoon
seasons,
the
top
two
are
drought
years.
The
rainfall
totals
are
on
the
right
and
the
third
down
is
a
flood
year,
and
so
you
can
see
that
the
the
short-term
variability
which
is
associated
with
the
miso
is
important
to
both
the
short-term
and
seasonal
term.
K
So
I'll
just
start
by
saying
that
to
identify
the
signature
of
the
in
rainfall
data,
we
used
a
method
based
on
the
work
of
suelos
and
collaborators
collaborators
in
2013..
K
K
So
this
sort
of
bimodal
north-south
distribution
of
rainfall
is
what
we're
looking
for.
But
it's
also,
you
know,
favored
by
the
empirical
orthogonal
function,
method
itself,
but
all
the
models
consistently
get
us
the
the
wider,
east-west
distribution
of
rainfall.
That's
consistent
with
the
re-analysis
data.
K
K
K
But
if
we
look
at
the
essentially,
if
you
look
at
the
distribution
of
phase
lengths,
the
control
run,
which
is
the
second
line
down,
gets
the
distribution
of
between
10
to
60
days
and
a
pretty
tight
distribution,
which
is
most
similar
to
the
phase
lengths
you
see
in
the
re-analysis,
and
you
can
see
that,
as
you
turn
off
turbulence
parameterizations,
you
start
to
get
a
wider
distribution
and
you
start
to
get
unrealistically
long,
mesophases
the
point
of
going
to
the
time
scale
of
the
monsoon
itself.
In
some
cases.
K
We
also
looked
at
try.
F
K
K
Climatology
or
mimoc,
which
shows
that,
although
all
the
model
versions
differ
from
the
observations,
especially
in
the
winter,
the
model
version
with
both
lanyard
circulation
and
mesoscale
eddies,
so
the
control
run
is
not
actually
the
one
that
gets
closest
in
the
summer.
That's
the
blue
line.
K
K
Quality
in
the
region-
and
it
also
means
that
probably
other
factors-
other
differences
in
the
two
model
versions.
Behavior
is
responsible
for
the
mesophase.
K
Result
so
basically,
these
results
indicate
that
the
turbulence
parameterizations
improve
the
coupled
models,
representation
of
the
miso.
K
C
F
K
F
For
what
you're,
looking
at
that's
a
little
less
subject
to
bakeries
of
the
atmospheric
model
to
produce.
K
Yeah,
no,
we
did
think
of
that.
The
the
couple
factors
are
one.
We
were
most
interested
in
rainfall
itself.
True.
K
As
opposed
to
looking
for
another
proxy
for
model
behavior
but
you're
right
in
that
the
re-analysis
data
set
just
as
a
as
a
you
know
as
an
example,
it's
it's.
A
A
Is
it
really
the
whole
basin
scale
behavior
of
mixed
layer,
depth
and
sst?
I
mean
I
mean
obviously
you're
looking
at
teleconnections
onto
the
continent,
but
can
you
can
you?
Can
you
identify
with
more
locality
where,
where
the
ocean
simulation
is-
and
you
know
deficient
in
its
representation
of
ac
fluxes.
K
A
Is
there
like
a
preferred
latitude
band
for
the
initiation
and
is
that
an
area
where
we
might
look
in
a
more
focused
process
level
way.
K
Yeah
I
mean
that's
where
you
start.
I
should
show.
K
I
don't
know
which
figure
is
best
to
show
it
but
sort
of
the
where
you
get
these
the
quote-unquote
sort
of
negative,
the
negative
dipole
kind
of
down
in
the
indian
ocean.
That's
sort
of
the
southernmost
extent
of
the
bands
starting
and
moving
up.
K
That's
how
you
get
that.
But
to
me
that
was
one
of
the
weaknesses
yeah
because
since
you're,
looking
at
the
accuracy
you're
looking
at
the
comparing
the
you're,
comparing
the
motions
sort
of
you're,
comparing
this
something
in
motion,
as
opposed
to
comparing
behavior
at
points.
K
K
L
C
So
we've
got
frank:
baylor
has
an
answer
in
the
chat
for
frank
there.
I.
C
It
out
peter
sullivan
has
a
as
a
questions
go
ahead
peter
I
want
to
see
it.
E
Hi,
patrick,
it's
just
a
simple
naive
thing:
do
you
have
any
clues
as
to
what's
missing
what
what
makes
what
would
move
all
those
curves
from
csm
to
what
you
think
the
observations
are.
K
I
think
well,
one
clue
is
how
it
gets
so
much
worse
in
the
winter,
so
the
the
the
the
difference-
the
monsoon
we're
looking
at
summer
months
here,
but
I
think,
if
I
were
to
just
guess,
sort
of
the
the
clue
you're
seeing
is
that
since
it
gets
since
the
mix
layer,
depth
difference
gets
so
much
worse
in
the
winter.
K
K
C
K
But
I
can't
quantify
that
right
now.
E
Yeah,
the
the
all
the
turbulence
wants
to
make
it
deeper
yeah,
if
it,
the
horizontal
transport
of
something,
might
make
it
shull.
But
no.
I
was
just
curious
if
you
had
any
thoughts
about
that.
C
G
C
If
the
model
salinity
is
so
biased
that
that's
what
you're
seeing
in
this
figure,
do
you
have
any
comment
on
that.
K
Right
because
it's
it's,
you
know
I
remember
we
were
looking
at.
We
were
talking
about
whether
it's
a
definition
problem
like
you're,
saying
between
the
definition
that
the
reanalysis
set
uses
in
the
the
model
uses.
C
B
K
So
like
what
there's
a
there's
a
bias
but
which
bias
are
we
noticing
basically
yeah
yeah?
So
we
we
worked
on
that
and
couldn't
we
looked
at
multiple
sort
of
re-analysis
data
sets
to
compare
it
to
to
see
which
you
were
seeing
but
yeah.
I
agree
that
that
is
was
yeah.
That
was
something
we
were
thinking
about.
C
All
right:
well,
thanks
everyone
for
your
questions
and
patrick
for
your
talk.
I
think
we're
gonna
move
on
next
to
cheryl
harrison,
I'm
pretty
sure
I
saw
cheryl
there.
She
is
you
want
to
go
ahead
and
share
your
screen
and
cheryl's
gonna
talk
to
us
about
disentangling
the
impacts
of
npp
and
sst
on
global
fisheries
declined.
M
M
C
C
C
Do
I
have
access
to
that
drive
folder,
you
should
I'm
looking
for
the
link.
Now
I'm
asking
our
support
team
to
get
that
link.
M
Okay,
I
thought
I
had
it
fixed,
but
it
didn't
so
let
me
let
me
email
it
to
ian
and.
M
So,
while
ian's
getting
that
up,
I'm
going
to
be
something
completely
different
than
the
previous
two
talks,
which
is
a
report
on
the
fisheries
model.
In
our
comparison
project
schmidt,
a
number
of
papers
coming
out,
one
of
which
is
using
csm
data
to
force
this
model
exclusively.
So
I
thought
it
would
be
a
good
opportunity
to
tell
you
guys
all
about
what
we're
doing.
M
My
name
is
cheryl
harrison.
I
was
a
postdoc
at
in
car
in
the
ocean
section.
I
am
an
assistant
professor
at
the
university
of
texas,
rio
grande
valley,
and
I'm
moving
to
louisiana
state
university
in
baton
rouge
to
the
department
of
coastal
oceanography
ocean
and
coastal
science.
M
Great,
thank
you
so,
and
I'm
the
fish
smith
earth
system
model
coordinator
next
slide,
please.
M
So
the
outline
for
this
talk
is
a
motivation.
What
is
fish
map
talk
about
some
background
marine
ecosystem
models
or
fisheries
models
which
we
abbreviate
as
mem's?
M
M
M
Thank
you
ian.
These
are
the
coordinators
currently
for
fishman.
It's
an
international
consortium
and
the
current
leader
is
julia
brand
blanchard
from
university
of
tasmania,
who
is
transitioning
from
derek
titansworth
and
splits.
The
time
between
da
housing
and
cambridge.
Previously
liam
chung,
haiki,
laza
and
eric
galbraith
were
coordinators
and
you
might
recognize
colleen,
patrick.
Some
of
you
are
working
with
her
and
on
fishery
stuff
next
slide,
please.
M
So
one
of
the
questions
to
answer
is:
how
do
we
stimulate
marine
ecosystems
with
our
system
models?
Our
question
keith
lindsay
a
discussion
case
lindsay,
and
I
had
about
this-
is
that
they're
just.
M
M
So
here
are
two
different
kind
of
think
about
how
to
do
it
on
the
left
is
a
kind
of
trophodynamic
model
that
charlie
stock
did
that
has
evolved
into
a
different
model,
feisty
that
colleen
patrick
uses
and
basically
in
that
model
on
the
left,
you're
dividing
up
trees
into
pelagic
fish
and
merciful,
fish,
which
are
on
the
bottom
of
the
seafloor,
and
you
take
two
different
drivers
for
those:
the
zooplankton
production
and
the
detrital
flux
to
the
sea
floor
and
then
there's
some
kind
of
transfer
efficiency.
M
That
says
how
do
we
get
from
these
food
sources
through
a
couple
different
phases
of
the
food
web
to
fish
and
that
transfer
efficiency
depends
on
temperature
and
other
things.
Another
way
to
look
at
the
models
is
through
the
size
structure
or
the
size
classes.
So
we
think
about
if
we
plot
on
the
right.
M
This
is
from
a
paper
by
ryan
hennigan,
one
of
the
earth's
one
of
the
coordinators
for
fishmap
who's,
a
mathematical
biologist,
and
if
we
plot
body
mass
size
class
versus
the
abundance,
there's
a
lot
more
things
than
there
are
large
things
and
there's
some
transfer
of
energy
or
carbon
biomass
from
small
things
to
large
things,
and
when
they
have
babies,
you
they
make
small
things
and
all
of
these
have
different
kinds
of
rules
associated
with
them
in
terms
of
growth
and
how
they
eat
each
other
and
how
they
reproduce,
and
that
depends
on
temperature
and
how
much
food
there
is
and
lots
of
different
things.
M
These
models
are
forced
by
net
primary
productivity
or
the
carbon
fixation
of
primary
producers,
phytoplankton
or
they're
forced
by
biomass,
biomass,
phytoplankton
or
biomass
of
zooplankton
and
and
we're
gonna
see
that
it
matters
a
lot
what's
forcing
them.
What
the
results
are
next
slide,
please.
M
So
how
does
climate
change
impact
bringing
ecosystems?
So
this
paper
by
ryan
is
all
about
that
through
the
lens
of
the
size
spectrum,
and
you
can
see,
we
have
the
slope
of
the
sizes,
so
body
mass
size
versus
abundance
and
one
of
the
things
that
we
think
happens
with
temperature
change
is
that
biomass
goes
down
all
over
and
this
is
a
physiological
effect
of
cold-blooded
species
how
they
work.
We
can
also
see
this
results
in
less
large
marine
animals
like
fish
and
sharks.
M
We
also
can
see
rain
shifts,
changes
in
spawning
tiny
timing,
miss
matches
and
predator
prey
relationships.
You
know
the
zooplankton
biomass
is
happening
at
a
different
time
than
when
all
the
fish
babies
are
happening,
and
so
they
starve.
This
is
something
that
has
happened
to
salmon
species
in
the
california
current,
for
example.
M
You
also
have
things
like
coral
bleaching,
habitat
and
for
coastal
species.
Sea
level
rise
can
be
a
big
problem
with
that
and
it's
a
complex,
non-linear
system.
You
know,
so
you
can
expect
that
the
unexpected
will
happen
and
there
will
be
weird
cascades
and
so
lots
lots
to
do.
Okay
next
slide,
please.
M
So
this
is
a
a
quick
summary
of
the
cmip
five
fish
mips
results,
and
this
isn't
a
paper
by
laza
at
all,
2019
in
pnas
and
the
the
title
tells
you.
The
big
take-home
message
here
is
that
when
you
have
climate
change,
you
get
a
trophic
amplification
trophic
amplification
means
that
larger
things
feel
the
impact
of
climate
change
more
than
smaller
things.
M
So
plots
a
and
b
here
on
the
bottom
left
are
the
change
in
the
ecosystem
biomass.
So
these
larger
animals
in
term
in
in
the
different
rcps.
So
we
have
on
the
bottom
panel
a
the
historical
is
the
black
and
then
it
breaks
into
the
four
different
research
concentration
pathways
or
carbon
co2
in
the
atmosphere
amounts.
I'm
sure
most
of
you
are
very
familiar
with
this
and
then
on.
The
right
is
a
breakdown
of
the
different
models.
M
I
think
it
was
the
opposite.
Vice
versa
and
there
are
six
global
marine
ecosystem
models,
and
so
what
we
see
here
is
a
decreasing
fish
biomass
over
time,
there's
a
large
spread
in
the
marine
ecosystem
models
and
there's
a
significant
difference
between
the
resource
concentration,
packed
with
rcps
and
so
yeah
next
slide.
M
So
this
is
another
set
of
figures
from
that
paper,
and
so
the
left
figure
shows
that
it
shows
four
different
lines:
green
is
not
primary
productivity
and
the
dark
green
is
phytoplankton,
the
yellow
is
zooplankton,
and
then
the
blue
is
fisheries,
and
what
we
can
see
here
is,
as
we
have
the
rcps
as
there's
more
warming,
more
co2.
M
We
get
decreasing
in
all
those
categories
of
the
marine
ecosystem,
but
we
have
a
larger
decrease
in
fish
than
we
do
in
the
lower
trophic
levels.
So
this
is
the
trophic
amplification,
the
bigger
the
organism,
the
more
the
impact
and
one
of
the
cool
things.
I
think
the
coolest
figure
from
the
most
interesting
figure
from
this
paper
is
if
we
plot
the
change
in
surface
air
temperature
over
all
the
rcps
and
the
change
in
total
biomass
at
the
end
of
the
century.
M
M
Next
slide,
please,
okay,
so
so
the
update
for
cement
six
for
those
of
you
who
haven't
been
following
or
along
cmipsic
in
general,
are
warmer
so
there's
higher
climate
sensitivity.
M
M
You
have
more
fisheries
decline
and
with
some
caveats
about
like
there's,
interesting
changes
in
the
drivers,
but
you
can
read
the
paper
if
you
really
want
to
know
about
that
slide,
please
again
so
more
warming,
you
have
greater
fisheries
decline,
so
the
solid
lines
here
show
c
map
six
and
the
dashed,
I
guess
they're
kind
of
dotted
lines
to
steam
at
five.
M
M
next
slide,
please,
okay,
so,
okay,
so
the
the
the
the
next
paper
I
want
to
talk
about
is
one
that's
also
in
review
in
progress
in
oceanography
right
now,
and
this
is
the
driver
decomposition.
So
it's
kind
of
a
brute
force,
taylor,
expansion
and
the
question
we
wanted
to
ask
is:
how
do
system
models
respond,
to
changes
in
temperature
and
lower
trophic
level
forcing?
And
so
we
just
divided
it
up.
M
We
forced
the
models
with
the
change
in
temperature,
but
kept
the
the
food
source
the
same,
and
then
it
did
the
future
change
in
food
source
and
temperature
exciting
and
then
saw
how
they
broke
down
and
I'll
have
a
slide
about
the
model
design
in
a
second.
So
just
some
background
here
to
really
think
is:
ecology
is
not
physics
and
exceptions
and
species
variation
rule.
You
have
a
need
to
open
up.
You
have
it
does
a
weird
thing,
and
so
there's
many
structural
and
there's
many
structural
decisions
to
make.
M
There's
a
question
that
I'm
gonna.
Sorry
it
popped
up.
I
got
distracted
I'll,
come
back
to
that
later,
so
there's
many
structural
decisions
to
make
and
how
to
go
from
our
system
models
to
fish
and-
and
you
know
you
can
think
of
that
as
like.
Oh,
it's
all
wrong
or,
but
it's
actually
really
interesting,
because
these
are
well-informed
decisions
about
how
these
biological
equations
work
and
so
that
variation
in
how
these
models
are
constructed
is
is
actually
really
interesting.
M
If
it's
warmer,
you
have
less
biomass
because
they're,
you
know
cold-blooded
animals
and
there,
but
there's
a
lot
of
uncertainties
in
the
structure
of
the
food
mobile
models,
including
the
details
of
how
these
temperature
impacts
vary
across
different
species
groups,
sizes
growth
in
metabolism,
how
to
simulate
the
food
transfer
exactly
how
climate
change
affects
that
transfer
and,
like
I
said
before,
they
take
two
different
various
trophic
level,
forcing
either
then
that
primary
productivity,
which
is
this
transfer
by
phytoplankton
from
inorganic
to
organic
carbon.
M
And
the
important
thing
to
note
here
is
that
primary
productivity
and
biomass
of
plankton
have
very
different
spatio-temporal
distributions,
and
I
highlighted
this
in
a
recent
paper
where
we
were
constructing
a
food
source
for
baby
turtles.
You
know
you
can
have
a
lot
of
primary
production
that
doesn't
translate
into
biomass
sticking
around
next
slide.
Please,
okay!
M
So
here's
the
here's,
the
model
design
we're
using
csm
cement
five,
and
that
was
because
I
dug
through
the
archive
and
found
all
of
the
files
we
needed
and
and
we
ran
in
ecosystem
models
in
the
high
emission
scenario
broke
it
down.
M
We
have
a
control
experiment
where
we
use
the
pre-industrial
control
for
everything,
and
then
we
have
a
temperature
change
where
we
we
just
change
the
temperature
and
a
lower
traffic
level
change
or
we
just
change
the
lower
trophic
level,
drivers,
whatever
they
are,
leave
their
same
and
then
all
climate
change,
and
that
all
climate
change
also
includes
some
of
the
models,
use,
oxygen
or
ph,
and
so
it
includes
all
those
other
drivers.
So
we
can
get
kind
of
a
residual.
It
doesn't
depend
on
this
linear
combination
of
temperature
and
food
source
for
for
marine
ecosystems.
M
So
this
is
a
slide
that
shows
the
different
drivers
for
the
csm
single
realization.
That's
the
cmip5
and
on
the
left,
are
the
the
levels
of
temperature,
primary
production,
phytoplankton
carbon
and
zooplankton
carbon
in
1950
to
1960
so
kind
of
pre-warming,
and
then
on.
The
right
are
the
percent
change,
and
so
we
can
see
that,
for
example,
if
you
look
at
the
bottom,
two
right
plots:
they
they
have
very
similar
spatio
temporal
distributions
on
the
whole.
M
But
if
we
look
at
knight
net
prime
reaction,
which
is
the
second,
I
wish
I
had
a
pointer
right
now,
and
can
you
point
to
panel
d?
Please
thank
you.
In
ld,
we
see
the
net
primary
production
chain
and
it
looks
really
different
from
the
biomass
change
and
so
shout
out
to
kristen
crumhert
who's
on
the
call
that
she
has
a
really
great
paper
talking
about
why
this
is
and
one
of
the
reasons
is
you
have
more
nutrient
recycling.
M
Sorry
there's
a
shift.
Sorry,
there's
a
shift
to
small
phytoplankton
by
atom
production
and
that
results
in
more
cycling.
So
you
you
have
you
just
have
really
different
different
you
don't
the
biomass
doesn't
stick
around
in
in
some
of
the
some
of
the
shifts.
Sorry,
I'm
not
explaining
this
well,
but
you
can
go,
read
kristin's
paper
and
she
tells
you
all
about
it
next,
please,
okay,
so
this
is
a
really
busy
slide
and
my
time
here
I
think,
I'm
out
of
I
think
I'm
out
of
time.
M
So
so
the
take
home
message
here
is
that
the
temperature
effects
have
really
large
uncertainty.
M
So
if
you
look
at
the
first
column,
looking
at
the
temperature
change,
it
really
really
varies
between
the
models
and
then
the
low
level
effects
agree
more,
but
it
really
depends
on
the
driver,
whether
they're,
using
a
mass
or
not
primary
productivity.
Next
slide,
please
and
again.
If
we
look
at
the
global
integrals
of
those,
we
can
see
the
driver
and
the
really
large
spread
in
the
temperature
change
impacts,
and
then
we
have
some
residuals
which
are
pretty
small
except
for
one
model.
So
mostly
it's
a
linear
combination
of
these
two.
M
These
the
the
temperature
and
lower
traffic
level
change
next
slide,
please
so
so
this
is
just
repeating
what
I
just
said,
and
so
next
slide
please
for
sake
of
time.
M
So
other
studies
we
have
in
the
works
for
fishermen
because
we're
comparing
the
regional
and
global
marine
ecosystem
models,
they
don't
agree
and
we're
doing
fishery
scenario,
development
to
go
with
the
ss.
So
how?
How
will
fishing
change
response
to
climate
change?
My
lab
is
working
on
climate
shocks
to
food
supply.
Kim
cher
will
give
giving
a
talk
this
afternoon
on
fisheries
and
packs
of
newport
in
the
climate,
variability
and
change
group
and,
and
then
fishmap
is
open
to
collaborators.
So
we
have
all
this
data.
M
C
I
probably
started
a
little
late,
so
I'm
gonna
keep
wrangling
the
slides
here
during
questions.
Frank,
could
you
could
you
maybe
watch
the
chat
window,
it's
kind
of
hard
for
me
to
keep
an
eye
on
that.
M
So
yeah
so
claudia
asks
suppose
we
would
not
warm
the
earth
but
cool
it
by
a
similar
rate.
What
would
that
do
to
fisheries?
Is
the
warming
the
problem,
or
rather
the
fact
that
the
system
changes
ecosystem
to
adopt
warming?
Is
the
problem
changes
in
food
supply
and
cool
it?
You
can
come
to
kim's
talk
because
new
does
cool
it
and
there
there
are
some
scenarios
that
we
do
for
nuclear
winter,
which
are
really
similar
to
volcanic
eruptions
which
pull
the
planet.
And
it's
it's
interesting.
M
It's
it's
a
little
different,
so
I'll
refer
that
to
there
and
these
ecosystems
in
these
models,
don't
really
adapt
they're,
not
that
sophisticated,
which
might
be
one
of
the
problems.
Is
that
we're
not
estimating
you
know
just
that
fish
can't
adapt
to
these
changes
in
in
their
habitat
really
quickly,
but
there's
a
lot
to
unpack
there,
and
so
I
I
I
would
love
to
discuss
it
more
with
you.
Thank
you.
A
I
had
a
quick
question,
so
I
guess,
as
you're
probably
aware,
matt
long
and
colleen
and
others
are
looking
at
coupling
some
of
one
or
more
of
these
fish
models.
Feisty
in
particular,
I
think
online
with
cesm.
Do
your
results
suggest
that
that
is
necessary
or
what?
What
advantage
would
there
be
to
having
an
online
coupled
versus
an
offline
driver
scenario?.
M
Yeah,
so
one
of
the
kind
of
discussion
points
for
this
paper
is:
there's
no
feedback
between
predators
and
lower
trophic
levels.
So
basically,
these
fish
have
unlimited
food
supply
when
you
force
them
offline
and
so
we're
missing.
These
key
processes
is
that
you
know
predation
does
change
jupiter
plankton,
which
changes
phytoplankton,
which
changes
biochemical
cycling
so
and
there's
higher
feeding
rates
are
expected
for
warming.
So
yes,
that
will
be
very
interesting
and
is
important.
M
A
First,
six
minutes
into
our
break,
so
I
think
we
should
go
ahead
and
take
a
break.
The
call
will
stay
open
for
anyone
who
wants
to
just
hang
out
and
chat
and
have
your
coffee
and
we
will
resume
at
10
20,
with
our
more
panel
like
session
on
wind
waves,
so
I'll
see
everybody
then.
C
C
M
C
Right
back
when
yeah
a
couple
of
things,
you
know
david
trostman,
yeah,
yeah,.
M
C
He's
great
and
then
I
also
I've
been
working
with
well.
Hillary
schnell
just
finished
with
me,
and
you
know
she
works
on
marine
heat
waves,
and
so
I've
been
thinking
a
lot
about
that.
So
I'm
pretty
interested
in
your
work
and-
and
this
mep
too,
because
of
you
know,
the
justification
for
studying
marine
heatwaves
is
always
fisheries
and
yet
I
know
nothing
about
fisheries,
so
I
just
wanted
to
make
that
connection,
because
I
think
it's
really
important
yeah.
M
And
and
that's
you
know
exactly
the
kind
of
thing
like
if
you
had
a
student
who
wanted
to
interrogate
these
models
looking
at
marine
heat
waves
that
are
resolved,
it
would
be
really
cool
to
bring
those
physical
forcings
and
their
run.
You
know,
so
we
have
this
data
and
we're
really
keen
to
to
share
it
and
for
cool
stuff
to
come
out
of
it.
Yeah
yeah
yeah.
C
C
She
developed
a
marine
heatwave
tracker
to
track
them
in
space
and
time
and
then
and
the
first
step
of
this
project
is
to
apply
that
to
one
of
the
one
of
the
large
ensemble
simulations
and
then
to
because
we
she's
done
it
on
oisst,
but
to
see
what
we
can
see
in
the
large
ensemble
and
if
it
kind
of
works
in
a
similar
way
or
not.
M
M
You
know
the
warming
hole
yeah
and
how
that
might
impact
fisheries,
and
I
think
we
might
have
convinced
some
people
to
run
some
fisheries
models
on
and
on
ensemble,
which
has
not
been
done
and
so
yeah
like
so
people
have
ideas
about
that
and
have
students
who
want
to.
Is
that,
like,
I
think
we
could
talk
people
into
running
fisheries
models
on
ensembles
pretty
easily.
C
So
well
we'll.
N
So
I
wanted
cheryl,
I
mean
so
many
of
the
things
you
guys
have
just
been
saying
I
I
would
love
to
to
stay
in
the
loop
on
so
marine
heat
waves
have
been
known
to
kill
off
the
eelgrass
in
the
chesapeake
bay
estuary
system,
also
in
our
back
barrier
system
here
in
virginia,
probably
elsewhere
too,
and
so
I
have
an
interest
in
that.
I
don't
know
cheryl
if
your
ecosystem
models
have
things
other
than
fish
in
them,
but
I
think
that
that
a
connection
could
be
made.
N
I
also
have
colleagues
who
are
working
in
fisheries
and
I'm
currently
part
of
a
team
looking
at
impacts
of
offshore
wind,
but
part
of
that
is
detangling,
what's
happening
with
offshore
wind
and
climate
change
and
we're
looking
at
the
impacts
on
fisheries
and
we've
already
seen
like
the
summer.
Flounder
used
to
be
off
the
coast
of
virginia,
and
now
our
fishermen
are
having
to
go
up
to
new
england
to
catch
the
flounder
so
anyway,
I
think
there
are
a
lot
of
connections
that
we
can
make
here
too.
So
I'd
I'd
love.
N
Yeah-
and
we
have
various
different
tools
like
one
of
them-
is
a-
I
think
it's
noah's
coastal
tool,
and
then
I
manage
a
coastal
resilience
tool
where,
like
you
know,
there
could
be
actual
application
of
some
of
the
things
that
that
we
learn
fabulous.
M
That
sounds
really
cool.
Isn't
we
don't
the
earth
system
model
like
these
global
ones?
Don't
resolve
the
coast?
Well,
the
earth
system
models,
don't
resolve
the
coast.
Well
and
the
you
know
they
don't
resolve
chesapeake
bay
for
sure,
but
there
are
a
lot
of
regional
models
and
we
do
have
a
lot
of
regional
fisheries
models
in
fish
and
I
can
connect
you
with
the
people
coordinating
that.
M
M
M
Yeah,
so
those
are
global
models,
and
but
there
are
a
bunch
of
regional
models,
regional
fisheries
models
that
take
that
work
in
various
ways,
yeah,
and
it
would
be
great
for
you
to
connect
with
those
people.
So
I
would
recommend
I'm
happy
to
talk
to
you,
but
I
would
recommend
talking
to
tyler
eddie
who's
up
in
newfoundland,
newfoundland
he's
and
and
derek
tinsoar
is
in
dalhousie,
so
they're
on,
you
know
coast
as
you
and
with
similar
currents
that
are
connecting
you
all
so.
C
C
No,
it
doesn't,
but
at
least
that
is
it
is
10
kilometers
so
that
it
probably
doesn't
do
the
chesapeake
or
puget
sound
very
well,
but
at
least
there's
going
to
be
more
physical
variables,
available
sort
of
for
the
climate
change
simulations
and
stuff.
What.
M
C
M
M
On
full
yeah,
so
there's
there's
a
lot
of
fun
things
we
can
do
with
that.
That's
that's
a
rich
database
to
play
around
with
we're.
Also
there's
you
know
the
couple
of
these
high
resolution.
Kind
of
hind
cast
analyses.
Csm
has
one
and
gftl
has
one
and
we
are
fishmap
is
planning
to
do.
C
Yeah
and
I
didn't
think
I'd
I'd
get
new
connections
by
logging
into
a
zoom
meeting,
so
there
you
go.
I
mean
I
know
you
cheryl,
but
still
to
reestablish
those
connections.
It's
great.
C
C
We're
doing
some
interesting
work
in
antarctica.
It
just
made
me
think,
and
kristen's
kirsten's
really
involved
with
that,
but
we're
looking
at
like
influence
of
pelinas
and
open
water
on
predators
and
predator
populations
and
then
alice
davier
just
got
a
proposal
funded
to
work
with
to
start
using
the
model
and
the
biology
and
stuff
to
try
to
make
recommendations
for
marine
protected
areas.
M
C
Yeah
and
in
antarctica
it's
it's
just
complicated,
because
it's
all
these
different
countries
as
well,
so
each
country
gets
to
have
all
the
you
know.
It's
like,
oh
my
god,
or
so
I've
heard
so
we're
just
we're
still
going
to
just
kind
of
plot,
along
and
and
and
also
this
idea
that
the
best
areas
to
protect
might
change
in
time.
So
as
the
climate
changes
that
the
area
that
you
thought
was
was
would
provide
the
biggest
benefit
in
protecting
that
may
change
geographically.
M
That
yeah,
that's
really
interesting
yeah.
That's
really
interesting,
I
feel
like
we
need
to
have
like
a
special
session.
I
guess,
like
ocean
science
is
just
happening
like
you
know,
using
earth
system
models
for
marine
ecosystems.
Maybe
there
is
going
to
be
a
session
about
this.
That
would
be
a
great.
It
would
be.
N
M
C
M
M
Okay,
fast,
laura
we're
also
making
a
lot
of
progress
on
the
antarctic
ice
nuclear
war
paper.
We'll
tell
you
about.
Oh.
C
M
A
C
A
So
please
take
advantage
of
that
all
right,
so
we're
now
going
to
launch
into
a
session
on
wind
waves
and
we've
sort
of
set
this
up
more
like
a
panel
discussion.
So
we
have
five
six
panelists
each
will
give
about
a
10
10
to
12
minute
presentation
on
their
perspective
on
where
we
should
be
going,
and
then
my
intention
and
hope
is
that
those
panelists
will
then
lead
the
discussion
because
they're
the
ones
that
know
where
what
the
issues
are
and
what
the
questions
are
and
what
the
technical
challenges
are.
A
So
I'm
hoping
that
they
will
help
not
help,
but
we'll
we'll
carry
that
discussion
forward
and
once
we've
gotten
through
the
talks
too,
we
have
as
an
experiment.
A
A
Webpage
is
sort
of
a
a
a
marker
of
where
you
know
where
we
got
to
today
that
people
can
look
back
at
and
refer
to
and
other
groups
working
with,
mom
6
can
sort
of
see
where
we
as
a
project,
are
going
with
this
and
this
the
last
point
is:
we
very
much
want
this
to
be
a
an
effort
that
benefits
the
entire
mom
six
developer
community.
A
G
Sure
so,
hello,
everybody,
I
we've
been
messing
around
with
waves
and
kind
of
an
experimental
context
for
a
bunch
of
years.
I
wanted
to
talk
about
where
it
looks
like
things
are
headed
now,
which
has
a
bunch
of
different
complexities.
To
think
about.
G
One
key
part
is
the
wave
driven
forcing
itself,
and
the
second
part
is
how
it
breaks
down
by
scale.
So
whatever
this
is
a
nice
movie
of
waves,
so,
as
lots
of
you
have
probably
seen
before
large
eddy
simulations,
the
kind
of
peter
sullivan
and
jim
mcwilliams
really
pioneered
can
simulate
the
turbulence
response
to
waves,
and
this
is
a
nice
example.
The
colors
are
temperature.
G
G
Ching
li
and
other
students
in
my
group
have
really
led
an
effort
to
show
that
we
can
significantly
reduce
the
mixed
layer,
depth
biases
in
most
seasons
and
most
of
the
world
by
including
that
small
scale
mixing
driven
by
waves
and
it
does
more
mixing,
and
it
also
does
more
entrainment
into
the
mixed
layer
I'll
skip
over
that
little
flame
effect.
G
And
then
ching
has
also
shown
that
using
wave
watch
may
not
be
necessary.
If
that's,
all
we
want
to
do
is
to
add
that
extra
little
bit
of
mixing
and
he
developed
in
collaboration
with
ivan
breivik,
who
and
adrian
webb,
who
are
really
on
the
wave
side
of
stuff
a
statistical
model.
G
So
what
we
call
theory
waves,
which
is
just
a
prediction
of
what
the
wave
field
would
be
like
in
a
model
like
wave
watch
without
simulating
the
waves,
and
so
you
can
see
if
you
can
jump
back
and
forth
the
lower
right
hand
panel.
There
is
showing
that
the
mixed
layer
depth
is
essentially
almost
indistinguishable
between
when
we
use
the
theory
waves
version
and
when
we
use
the
full
wave
watch.
G
G
G
G
Actually,
you
can
see
their
effects
the
effect
of
this
stoke
shear
force
on
the
downward
velocity,
and
so
it
looks
like
a
langmuir
cell
one
of
these
typical
language
cells
out,
but
just
a
really
big
one.
So
apparently
the
same
phenomena
is
happening
on
mesoscales
and
sub
as
the
scales,
as
is
happening
in
the
language
scale,
but,
of
course,
with
different
scales.
G
Oh-
and
this
is
the
last
aspect-
so
sean
haney
in
his
phd
thesis
showed
that
it's
really
important
to
consider
the
so-called
anti-stokes
effect,
which
is
where
the
eulerian
current
responds
to
oppose
the
stokes
drift.
I'm
not
going
to
get
into
details
of
that,
but
brody
pearson
wrote
a
very
nice
paper
on
how
difficult
it
is
to
think
about
this
and
how
even
large
eddy
simulations
are
probably
doing.
G
This
effect
a
little
bit
wrong:
okay,
so
nobu
suzuki-
and
I
thought
a
lot
about
this
set
of
equations
and
how
to
make
it
simpler
to
understand-
and
one
simple
thing
to
do
is
which
was
shown
before
by
darrell
holm,
but
we
kind
of
wrote
the
equations
in
a
cleaner,
more
approachable
format.
G
It's
just
one
way
to
think
about
this
is
that
advection
of
everything,
invention
of
momentum,
of
action
of
buoyancy,
advection
of
salinity,
erection
of
vorticity
and
tilting
of
stretching
of
vorticity
infection
of
potential
vorticity
advaction
of
t
turbulent
kinetic
energy
can
all
be
thought
of
as
affected
by
the
lagrangian
velocity,
and
that's
one
of
the
big
effects
of
lagrangian.
Here
is
the
stokes
plus
eulerian
velocity,
and
then
these
other
two
forces
you
have
to
consider
in
the
momentum
budget,
the
lagrangian
coriolis
effect
and
the
stoke
shear
force,
which
I
already
mentioned.
G
If
you
have
those
three
effects,
that's
actually
everything
you
need
in
a
large
eddy
simulation
and
we
spent
some
extra
time
thinking
about
what
how
you
could
do
this
in
a
climate
model
as
well
to
get
maybe
those
front
frontal
acceleration
pieces
that
I
talked
about.
Even
in
a
high
resolution,
climate
model,
which
is
ahead
of
what
we're
doing
maybe
now
but
that's
cool-
to
have
something
like
mom
six
and
then
the
other
piece
of
the
story
is.
We
can't
really
do
this
lagrangian
advection
for
everything.
G
Unless
we
have
the
lagrangian
coriolis
force
there
present
to
make
the
correct
anti-stokes
effect.
A
lot
of
groups
have
experimented
with
just
adding
stokes
drift
on
top
of
whatever
the
model
velocity
is
and
then
seeing
what
it
does
to
things
like
surface
drifters.
That's
actually
dynamically
incorrect,
because
what
happens
when
you
add
in
these
lagrangian
coriolis
and
stoke
shear
forces,
is
that
it
changes
the
velocity
field,
and
then
you
have
to
re-evaluate
and
go
backwards
to
figure
out
what
the
objection
is
really
doing
to
build.
G
This
whole
set
consistently,
okay,
so,
but
to
the
main
point
here.
So
this
is
the
set
of
equations
that
nobu
and
I
studied,
and
so
here's
that
stoke
shear
force.
So
this
is
the
normal
eulerian
flow
and
not
including
wave
effects.
If
we
include
the
wave
effects,
it
just
makes
a
lot
of
lagrangian
velocities
appear
and
the
stoke
shear
force
appear.
G
You
can
do
some
algebra
and
bing
bing
bing,
expand
it
out
then
make
the
hydrostatic
approximation,
and
this
set
at
the
bottom
is
totally
consistent,
being
included
in
a
model
like
mom
6,
and
that
would
be
easy
to
do.
G
But
when
brandon
got
ex
inspired
to
put
this
into
mom
six,
he
wanted
to
use
the
stokes
a
the
so-called
vector
invariant
form
of
mom
six,
which
looks
like
this
at
the
top.
So
it
doesn't
have
an
eviction
term.
So
it's
not
so
obvious
where
you
stick,
the
lagrangians,
but
daryl
holm
had
shown
in
the
past
and
then
nobu,
and
I
revisited
to
show
that
this
is
the
correct
form
and
that
in
that
form,
which,
if
you
expand
out,
then
make
the
hydrostatic
approximation.
G
So
the
point
here
is
is
that,
despite
the
fact
that
these
two
equations
look
quite
different,
they
end
up
in
the
same
place.
Even
surviving
making
the
hydrostatic
approximation,
even
though
the
wavy
hydrostatic
model
isn't
really
hydrostatic.
It's
got
a
wave
impact
in
the
vertical
momentum
equation
and
the
amount
I
should
have
said
of
how
big
this
term
is
versus.
This
term
is
what
that
epsilon
parameter
is
really
governing.
G
So
epsilon
is
a
way
of
checking
on
that
and
epsilon
is
typically
pretty
small,
for
course,
resolution
models,
but
it
may
be
as
big
as
or
bigger
than
the
rossby
number,
and
so
we
definitely
are
very
concerned
with
things
that
are
ageostrophic
transports.
So
I
guess
equivalently.
We
should
be
think
thinking
about
things
that
are
non-hydrostatic
effects
or
wavy
hydrostatic
effects
that
are
equivalent
small
tweaks
to
the
model
system,
even
at
coarse
resolution.
G
Now.
The
last
part
of
the
map
that
I
wanted
to
highlight
here
is
whether
we
write
it
in
the
way
that
the
vector
invariant
form
or
in
the
invection
form
you'll
note
that
these
wave
equations
only
appear
in
either
linear
terms,
if
you're
kind
of
thinking
about
distributing
them
out
or
in
nonlinear
terms,
with
itself
so
stokes
drift
times,
stokes
drift
and
in
either
case
whether
it
appears
linear,
in
which
case,
there's
no
scale
to
scale
non-linearity
transfer
or
stokes
times.
G
That's
a
large
scale
times
a
large
scale
correlation,
and
so,
when
you
think
about
it,
transferring
from
the
large
scale
to
other
scales,
it
doesn't
go
very
far.
It
actually
can't
jump
very
many
wave
numbers.
G
So
what
this
means
is
is
that
we
won't
end
up
double
counting
if
we
are
parameterizing
small-scale
stokes
drift
effects,
at
the
same
time
as
resolving
large
scale,
stokes
drift
effects,
those
we
don't
get
a
cross-scale
coupling
the
way
you
would
with
normal
turbulence
problems
because
of
the
way
that
the
stokes
drift
appears
in
these
equations
and
the
way
that
the
wave
averaged
equations
work
themselves.
G
And
so
it
looked
like
this,
where
wave
watch
was
talking
with
the
coupler
and
being
passed
over
to
pop
we're
presently
now
in
development,
with
a
bunch
of
other
groups,
building
a
new
opsi
cap
to
talk
to
the
new
ops
mediator,
which
allows
an
upgrade
of
the
version
of
wave
watch,
but
also
allows
us
to
link
up
to
mom
6,
and
we
are
about
to
be
finished
with
doing
a
theory.
Waves
version,
which
replaces
wave
watch
through
the
same
neuopsy
mediator
system
and
that
will
be
make
it
much
cheaper.
G
We
watch
at
present
we're
running
about
three
by
four
degrees
resolution
and
cesm2
to
go
to
one
degree
to
match.
The
wave
grid
to
the
ocean
grid
is
about
60
times
slower
and
to
go
to
the
10th
degree
to
match
the
high
resolution
ocean
models
about
60
000
times
slower,
whereas
the
theory
waves
have
a
negligible
cost
at
all
resolutions.
G
So
we'd
like
to
have
that
basically
for
efficiency,
even
though
we
know
the
accuracy
will
be
lower.
That's
all
I
wanted
to
say
so.
If
there
are
any
questions.
A
Okay,
I
think
what
the
plan
was
we're
gonna
unless
there's
a
question
that
requires
clarification.
Well.
Well,
that's
those
now,
but
we'll
hold
broader
questions
to
the
discussion
after
we've
had
each
of
the
panelists
speak.
A
So,
let's
move
on,
I
think
brandon
was
no
wait.
Let's
see
yeah
brandon.
O
Thanks
frank:
I'm
trying
to
do
this
through
the
web
interface,
so
hopefully
this
is
all
working
right.
Please,
let
me
know
if
not.
I
can't
tell
if
my
video
is
even
working.
O
O
All
right
excellent!
So
thanks
for
the
invitation
to
join
your
your
meeting
this
morning,
for
those
of
you
don't
know
me,
I'm
brandon
reichel,
I'm
at
gfdl
I've
been
working
on
incorporating
some
of
these
exact
same
stoke
strift
effects
that
baylor
was
just
talking
about
in
mom
six
for
about
a
half
decade.
O
Now
I'm
really
excited
that
this
is
being
undertaken
by
the
cesm
group
now,
so
I'm
excited
to
work
with
you
guys
on
this,
so
I
was
just
going
to
give
a
little
bit
of
an
overview
about
where
we
stand
right
now
with
what's
in
mom
6
and
what's
next
in
our
plans,
so
I
could
probably
skip
through
most
of
this,
but
just
a
basic
introduction.
What
is
stoke
strip
so
baylor
showed
the
nice
gif
image
that
shows
these
net
translation
of
fluid
particles.
O
As
you
follow
wave
orbitals,
it
gives
you
basically
a
vertically
decaying
current
drift,
that's
associated
with
the
surface
waves
that
can
interact
with
the
background
what
we
often
call
the
eulerian
current
through
these
equations.
That
biller
was
just
showing
one
way
to
calculate
the
stokes
drift
is.
If
you
have
a
wave
spectrum,
you
can
use
this
equation,
given
here
at
the
bottom.
The
wave
spectrum
here
is
given
by
psi
wave
number
k
direction,
theta
frequency
omega
and
then
the
depth,
of
course,
is
z.
O
You
could
use
this
relationship
from
the
wave
spectrum
to
get
the
stokes
drift,
the
horizontal
components
of
the
stokes
shift
at
some
depth
z,
if
you're
not
familiar
with
the
concept
of
a
wave
spectrum.
Basically,
you
can
take
a
wave
field,
something
like
this
image
from
a
picture.
Steve
griffey's
took
a
few
years
ago
in
the
southern
ocean
and
you
can
decompose
this
wave
field
into
the
various
direction
and
wave
number
weight,
inverse
wavelength
components.
So
that's
what's
showing
in
this
image
here
all
the
way
on
the
right.
O
This
happens
to
be
a
five
meter
wave
field
with
a
wind
velocity
of
wind
speed
of
12
meter
per
second.
The
vector
is
shown
by
this
blue
arrow
here
and
you
can
see
based
on
the
contours
where
the
wave
energy
happens
to
be
concentrated
in
this
instance.
So
you
can
actually
see
two
different
wave
fields
here:
one
that's
about
10,
maybe
15
degrees,
counterclockwise.
O
To
the
wind
and
one
the
other
direction
clockwise
to
the
wind
and
the
interesting
thing
about
this,
especially
because
these
are
happening
at
different
wave
numbers
or
wavelengths-
is
that
we
can
have
a
turning
of
the
stokes
drift
with
depth.
So
I'm
going
to
come
back
to
why
why
that's
relevant
in
a
bit?
Why
is
stokes
drift
important
in
an
ocean
model,
so
there's
kind
of
two
different
scales
of
how
we
think
about
this
again
bill?
I
covered
this
very
nicely
at
small
scales,
the
turbulent
scales.
O
We
have
significant
enhancement
of
these
vertical
fluxes,
so
these
are
the
large
eddy
simulations
again.
Baylor
was
just
talking
about
so
what
I
just
wanted
to
show.
These
are
some
results
from
some
specific,
less
run
by
ching
li
several
years
ago.
Everything's
identical
in
these
two
runs,
except
for
one
of
these
is
stokes
drift
and
one
doesn't
I'm
using
the
same
color
bar
here
to
show
you
the
turbulent
buoyancy
flux.
O
This
is
basically
showing
how
much
we're
in
training
the
you
know
the
colder
heavier
water
from
below
the
mix
layer
into
the
mix
layer,
and
you
can
see
significant
enhancement
in
this
run.
With
this,
the
stokes
drift
and,
and
basically
this
langmuir
turbulence
phenomenon,
and
so
we
want
to
include
these
langmuir
turbulence
effects
in
our
vertical
mixing,
parameterizations
we're
using
in
the
model
we
heard
about
that
a
little
bit
before
the
break.
So
this
comes
from
a
paper.
Probably
most
of
the
co-authors
are
on
this
talk.
O
Ching
led
this
back
in
2019,
comparing
different
schemes
with
langmuir
turbulence.
This
is
a
1d
simulation
at
ocean
station
papa
sorry,
the
x
axis
is
missing.
This
is
about
a
year
cycle
and
this
is
just
showing
the
evolution
of
buoyancy
in
the
mix
layer
in
a
standard
configuration.
So
this
is
just
a
standard,
kpp
model
through
cv
mix,
and
then
these
are
three
different
langmuir
turbulence,
parameterizations,
each
of
which
has
been
kind
of
separately
developed
from
large
eddy
simulation
results.
In
this
particular
case.
O
What
you
can
see
so
the
the
red
blue
color
scale
now
in
these
three
panels
is
the
difference
from
this
top
b.
You
know
top
left
b
panel
here
and
you
can
see
additional
entrainment
when
we
account
for
langmuir
turbulence
in
our
parameterizations,
and
it's
somewhat
similar
in
this
case
between
the
three
parameterizations.
That's
not
always
going
to
be
the
case,
and
I
would
highly
recommend
looking
at
this
paper
for
a
lot
more
details
of
the
discussion
of
kind
of
the
state
of
the
science
there
anyways
more
specific.
O
What
we
have
right
now
available
in
mom
6.
mom6
can
use
kpp
and
epbl
epbl's
the
boundary
layer
scheme
we've
been
using
at
gfdl.
O
Both
of
these
already
include
some
langmuir
turbulence
options,
so
some
of
the
schemes
that
baylor
was
talking
about
in
the
work
jing
had
done
as
well
as
then
the
work
we've
done
in
epbl.
This
is
just
showing
a
a
particular
snapshot
of
the
boundary
layer
depth.
So
this
is
the
summer
in
the
southern
hemisphere.
This
used
to
be
a
particularly
problematic
region.
In
these
models.
We
didn't
get
deep
enough
mixing
for
a
variety
of
reasons,
but
one
of
these
being
missing
the
the
langmuir
turbulence
effect.
O
So
from
these
two
locations,
I've
just
shown
the
median
annual
cycle
of
mixed
layer
depth
the
model
in
black
and
an
argo
based
metric
is
in
red.
You
can
see
we're
starting
to
do
a
pretty
good
job
of
resolving
the
annual
cycle
here,
including
getting
the
depth
in
the
summer
time
much
more
reasonable
than
perhaps
it
was
these
other
two.
So,
unfortunately,
I
haven't
run
this
model
without
langmuir
we've
been
pretty
happy
with
what
it's
done
for
us.
O
This
is
the
specific
year
2018
in
these
runs,
and
so
this
is
basically
an
enhancement
over
the
mixing
in
the
absence
of
language
somewhere
between
10
and
even
up
to
40
in
some
of
these.
So
this
is
just
the
mechanical
component.
You
also,
of
course,
have
convective
mixing.
So
what's
next
here,
of
course,
when
we
run
more
large
eddy
simulations,
we
might
be
coming
up
with
new
and
improved
mixing
parameterizations.
O
It's
a
bit
of
a
maybe
a
moving
target
in
that
sense,
but-
and
bill
might
talk
a
little
bit
about
this,
I
think
next,
and
we
also
can
kind
of
I've
been
thinking
about
incorporating
some
of
these
higher
order.
Turbulent
closure
schemes
such
as
the
second
moment,
miller,
yamada
type
schemes-
be
it
got
them,
then
skipping
ahead
to
the
to
the
larger
scales.
So
so
baylor
just
spent
a
lot
of
time
on
this.
O
O
Maybe
these
terms
don't
do
a
whole
lot,
but
it's
it's
best,
as
baylor
said,
to
keep
all
the
dynamics
consistent,
we
need
to
be
able
to
make
sure
if,
if
we
want
to
get
the
total
lagrangian
advection
that
we're
accounting
for
things
like
stokes,
coriolis
and
everything,
and
so
the
main
point
to
see
here-
these
are
two
two
sets
of
equation.
One
is
the
standard
equations
in
mom
six
right
now
the
vector
invariant
form,
which
is
important
for
mom.
O
Six
being
a
c
grid
model,
and
then
this
is
the
hydrostatic
relation
two.
So
this
is
the
horizontal
equation.
One
horizontal
momentum:
three
is
our
continuity
equation
in
the
generalized
coordinate
and
and
four
would
be
basically
a
tracer
equation.
O
So
theta
would
be
the
temperature
our
salinity
could
be
represented
by
s,
and
so
we've
been
working
on
adding
the
the
stokes
drift
via
this
kind
of
this
craig
lipovich
averaging
procedure
to
these
resolve
skill
terms-
and
I
basically
highlighted
and
read
the
same
terms
that
baylor
had
in
yellow
in
his
slides-
and
the
point
of
this
is
to
emphasize
that
we
can
actually
distill
this
down
to
just
needing
to
modify
these
specific
terms
in
the
model
to
get
the
right
transports.
O
Finally,
I
want
to
get
into
some
of
the
more
technical
sides
of
this
conversation,
so
this
is
how
we
actually
get
stoke
strip
into
mom
6
from
wave
watch
3..
As
I
mentioned,
we
can
use
a
formula
like
this
to
get
the
stoke
strip
profile
from
the
wave
spectrum.
What
we
really
want
is
we
want
the
stokes
drift
averaged
over
a
layer,
so
we
need
to
then
be
integrating
this
over
the
layer
depth
to
get
the
total
transport
and
then
convert
that
into
an
average.
O
The
full
spectrum
usually
of
wave
watch
simulation
does
something
like
25
wave
numbers
in
24
directions.
That's
a
lot
of
information
to
be
saving
or
passing
back
and
forth.
Now
we
only
need
two
directions:
the
x
and
y.
So
we
can
collapse
this
down
to
about
50
components.
That's
still
quite
a
bit,
so
the
strategy
we've
taken
is
to
reduce
this
from
25
wave
numbers
to
what
I'm
calling
here,
n
wave
numbers.
So
n
could
be.
You
know
any
integer
between
1
and
25.
O
Basically-
and
this
isn't
an
exact
relationship
that
can
be
derived
from
this,
but
it's
more
of
a
parameterization
using
the
idea
of
a
monochromatic
spectrum
where
we
have
a
surface
velocity
and
some
decay
scale
k
associated
with
that
particular
surface
velocity.
O
The
advantages
of
taking
this
approach
is
that
we
can
make
n
much
less
than
25
and
and
and
so
closely
mimic
having
the
full
spectrum.
This
exponential
form
lends
pretty
well
to
what
I
call
here
semi-analytically
evaluating
terms,
so
we
can
do
derivatives
and
integrals
quite
readily
over
exponential
functions,
which
is
nice
for
implementing
this
in
a
model
and
with
n
being
a
relatively
modest
analysis
on
the
next
slide,
we
can
basically
capture
the
broadband
effects
pretty
well
so
to
answer
the
question:
how
many
ends
do
we
really
need?
O
This
is
one
particular
case
here.
This
is
from
the
southern
ocean
in
in
a
jra
forced
simulation.
The
surface
stokes
drift
is
about
20
meter
per
second,
so
black
is
the
solution.
You'd
get
with
the
full
spectrum,
and
then
I've
taken
a
number
of
values
of
n
here,
so
moving
from
dark
red
to
dark
blue,
where
we
have
very
few
to
a
lot
of
numbers,
in
forgetting
the
magnitude
which
is
the
left
profile
here
turns
out,
we
probably
need
something
like
at
least
three
of
these.
You
know
stoke
bands.
O
Basically,
this
happens
to
be
a
case
where
the
stokes
drift
turns
a
lot
in
depth
and
with
only
three
components.
We
don't
quite
capture
that
at
the
deepest
part.
So
maybe
you
need
to
go
a
little
bit
more.
O
If
you
want
to
capture
that
depth
dependence
of
the
directionality
of
the
stokes
drift,
but
the
point
is
we
can
collapse
this
down
now
to
something
like
six
to
12
components
versus
50,
which
is
you
know,
a
nice
improvement,
whether
you're,
saving
or
passing
these
through
a
coupler
and
then
finally,
just
a
a
quick
recap
of
what
we've
done
for
using
stokes
drift
in
mam6
without
a
wave
model.
So
we
can
ingest
these
surface
stokes
bands
from
wave
watch
three
output
files.
This
is
already
in
the
wave
model.
O
The
latest
versions
of
the
wave
model
so
being
able
to
use
the
latest
version
in
csm
will
be
important
to
take
advantage
of
that
type
of
infrastructure.
This
is
kind
of
good
for
a
one-way
type
simulation
where
you
have
a
prescribed
atmosphere
within
epbl
and
kpp.
We
already
can
use
this
theory
wave
that
biller
just
talked
about
to
predict
the
langmuir
number,
which
helps
gives
us
the
language
turbulence
effect
on
the
mixing.
O
This
is
good
for
climate
models,
where
we
don't
have
the
resources
to
run
a
wave
model,
and
then
baylor
just
mentioned
a
bit
about
the
potential
for
using
theory
waves
then
to
actually
get
stokes
drift
profiles.
I
guess
that's
quite
a
bit
more
of
a
research
question
in
how
well
that's
going
to
work
and
perhaps
there's
other
simpler
approaches,
and
so
this
is
just
showing
you
know
some
of
these
ideas
for
how
we
get
the
stokes
drift
into
mom
six,
whether
it's
that
one
way
sort
of
storing
that
cdfs
two-way
coupling.
O
Whichever
cap
you
happen
to
use
these
are
at
varying
instances
of
completeness.
I
guess,
and
then
this
is
something
that's
maybe
more.
I
guess
it
sounds
like
baylor's
actually
pretty
close
on
this,
so
yeah,
that's
all
I
have.
I
look
forward
to
the
discussion
and
the
rest
of
the
talks
now
in
this
session.
A
Brandon
all
right
move
on
to
bill
large.
F
F
C
To
select
what
screen
you
want
to
share,
either
the
whole
screen
or
a
certain
window.
F
That
might
do
it,
that's
what
I
thought
I
was
doing
anyway.
Share
screen,
there's
the
there
they
got
it
now
there
we
go,
I'm
pushing
the
same
buttons,
okay,
so
I'm
going
to
just
follow
up
on
and
what
I
interpreted
much
of
the
discussions
that
you've
just
heard
over
the
last
month,
or
so
that's
about
all
we've
been
talking
about
this
a
lot.
The
ideas
develop
surface
waves
and
cesm
version
of
mom6,
and
what
we
need
to
do
today
is
to
develop
around
a
six-month
initial
implementation
plan.
F
The
goal
is
to
enable
research
on
surface
wave
effects
and
parameterizations,
the
thinking
being
when
you
hear
all
the
different
uses
that
these
models
are
being
put
to
in
coupled
mode.
Some
of
us
are
quite
worried
that
the
applications
are
out
stripping.
The
physics
that's
gone
into
it
and
model
development
needs
to
continue
if
we're
going
to
keep
using
it
in
all
these
various
and
sundry
ways,
which
you
know
we
want
to
have
some
confidence
that
those
are
the
best
answers
we
can
get.
F
But
here's
where
we
are
now
and
that's
a
good
basis
to
start
with
okay,
so
the
topics
I'll
talk
about
quickly
is
is
some
of
the
ocean
wave
coupling
details
and
what
we're
planning
to
do.
I've
underlined
alperl
tuntas,
who,
as
frank
mentioned
earlier,
we'll
talk
about
this
this
wave.
What
we've
done
with
wave
watch
three
at
the
beginning
of
the
discussion
and
it
answers
brandon's
question
of
what
what
we're
doing
on
that
and
these
recent
versions
and
things
then
there's
just
talking
about
the
momentum
equations.
F
That's
what
baylor's
been
talking
about
and
brandon
to
some
extent,
contributions
have
been
streamed
from
jim
mcwilliams
to
peter
sullivan,
and
then
someone
on
the
numerix,
mostly
coming
from
bob
haulberg,
with
conversations
with
brandon,
marcus
and
and
alper
and
chingly
just
quickly
the
experiments
we
now
have
our
alvara
and
gustavo
actually
went
in
and
and
got
enough
computing
time
to
run
a
number
of
single
gra
cycles
through
2018,
as
at
least
the
beginning
of
an
experimental
framework
of
how
we
might
test
these
various
numerics
that
we're
putting
in
to
make
sure
that
it's
available
for
people
to
do
their
their
research
on
these
wave
effects.
F
So
the
premise
we
have
is
that
there
will
only
be
one-way
ocean
coupling
to
a
hierarchy
of
ocean
waves.
You've
heard
something
about
the
wave
watch,
three
there's
also
her
theory
ways
or
statistical
waves.
We
also
used
to
have
data
models,
but
I'm
not
so
sure
that
the
theory
wave
might
just
take
the
place
of
the
data
waves,
but
it's
one-way
coupling
and
we
need
to
discuss
this
hierarchy,
which
which
of
these
options,
should
we
should
we
focus
on
to
get
into
the
malm
six
version
of
cesm3.
F
The
wave
model,
as
you
just
saw,
will
actually
go
or
send
to
the
coupler
or
mediator
on
the
wavegrid,
the
stoke
drift
component.
That
brandon
talked
about
where
he
told
me
and
was
about
going
to
be
about
six,
as
you
saw
now
the
difference.
Okay,
so
we
will
send
those
and
now
there's
the
question
of
what
are
the
other
requirements.
That's
part
of
the
discussion.
Nikki
talked
about
what
you
might
need
for
bubble,
gas
fluxes.
F
So
in
general,
there's
wave
dependent,
rc
flux
requirements.
What
other
requirements
we
have
some
legacy
stuff
to
think
about.
How
much
do
we
keep
bringing
in
from
before
so
that's
open
for
discussion,
but
what
I
would
think
about
who
who's
actually
gonna?
Do
it?
Who's
gonna
provide
the
scientific
guidance
when
and
justification
for
any
requirements
you
might
want
so
watch
it.
You
may
be
volunteering
yourself
if
you
step
up
and
say
you
want
something:
the
mediator
coupler.
F
F
Okay,
so
that
would
happen
in
the
mediator
coupler
as
far
as
we're
thinking
now
and
then
mom
6,
on
a
time
evolving,
spatially
variable
vertical
grid,
that's
possible
in
mom
6,
with
this
spatial
variability
of
the
grid.
These
are
the
stokes
profiles
that
brandon
mentioned
down
here,
and
the
lagrangian
are
just
as
some
as
you've
heard,
but
one
thing
I
did
notice
two
weeks
ago,
when
I
asked
brandon
this
decay
scale
was
fixed.
It
wasn't
a
function
of
the
particular
wave
monochromatic
wave
component
and
in
his
code
he's
got
that
with
an
index.
F
So
whatever
it
is,
it's
only
one
more
one
more
parameter,
I'm
not
sure
where
brandon
is
on
that,
but
whatever
it
is,
we
can
accommodate
and
that's
the
plan
to
stay
as
close
to
whatever
he's
he's
coming
up
with,
at
least
for
this
initial
implementation.
F
To
allow
people
to
get
started
on
these
experiments
that
they
might
want
to
do
as
far
as
the
momentum
equations
are
concerned,
the
ver
the
pr.
The
premise
is
that
our
vertical
resolutions
will
approach
a
meter
or
better,
ultimately,
we'd
like
to
really
just
get
a
nice
diurnal
cycle
of
the
diurnal
heating
layers.
Just
off
the
board.
You
know
just
resolved,
so
I
didn't
hear
anything
on
baylor
on
what
vertical
resolution.
So
I
hope
the
discussion
brings
up
some.
What
does
this
vertical
resolution
mean?
F
So
that's
gonna
require
one
would
think
mom
6
to
solve
a
form
of
the
same
wave,
average,
goose
and
x
equations
that
you've
just
been
looking
at
so
right
now,
if
we're
putting
these
parameterizations
based
on
las
into
a
model
that
doesn't
solve
the
same
equations,
you're
not
sure
you're
going
to
at
all
well
you're,
pretty
sure
you're
not
going
to
get
the
same
mean
fields
that
you
parameterize
with,
and
therefore
the
parameterizations
are
less
certain
your
uncertainty
rises.
F
So
that's
the
hope
that
these,
with
these
new
equations
that
you,
you
are
new
terms
in
the
equations
you've
seen
we'll
start
getting
the
the
flow
representative
in
the
in
in
the
gcm
that
we
get
in
the
las
the
additional
stokes
terms.
You've
heard
it
boy,
I
tell
you,
you
can
move
these
things
around
and
thank
you
baylor
for
sorting
out
a
lot
of
the
controversy
on
make
sure
you're
talking
about
the
same
equations.
F
When
you
start
talking
about
what
term
you've
got
so
you've
seen
all
those
and
we'll
just
take
what
brandon's
done
to
start
with
and
then
we'll
start
talking
about
the
vertical
flux
of
horizontal
momentum,
that's
the
stress,
vector
or
minus
the
stress
vector,
and
one
of
the
things
thinking
at
the
moment,
is
that
the
stress
and
shear
vectors
will
no
longer
need
to
be
aligned.
In
other
words,
we
won't
have
down
gradient
viscosity
baked
into
the
into
the
solution,
so
I'll
spend
some
time
on
that
particular
problem.
F
F
But
if
you
look
at
this
plot
over
here
of
the
actual
shear
aligned
with
the
wind,
this
is
relative
to
the
wind.
This
is
downwind.
This
is
crosswind
that
at
this
point
over
here,
it's
very
much
downwind,
but
it's
not
zero
in
the
crosswind,
and
so
that
means
that
the
shear,
I'm
just
guessing,
maybe
40
degrees,
to
the
right
of
the
the
wind,
in
this
case
at
four
meters
at
2
pm.
F
If
you
go
later
in
this
observational
record,
this
is
from
15
north
from
hughes
at
all,
and
these
are
rare
observations
I
must,
but
if
you
look
at
6
p.m,
at
14
meters
down
here,
there's
an
awful
lot
of
purple,
I.e,
negative
and
not
so
much
red,
so
maybe
you're
80
degrees
off
and
the
thought
is
what
the
heck's,
the
stress
vector
doing.
We
don't
have
measurements
of
the
stress
vector,
so
we
have
to
guess,
and
what
we've
been
saying
is
that
the
stress
is
aligned
with
the
shear.
F
That's
down
gradient
viscosity.
Well,
if
the
shear
is
you
know,
if
this
direction
of
the
shear
is
only
this
scale,
how
would
any
motions
in
those
big
heady
motions,
you've
seen
know
about
just
that
scale
when
they're
coming
from
above?
So
it's
not
at
all
obvious
that
the
stress
and
shear
are
aligned
when
the
stress
fat
or
the
shear
vector
has
such
small
vertical
scales
and
that's
just
some
observations.
We
have
and
we've
seen
other
places
of
observations
where
there's
small
vertical
scales
and
shear.
F
So
how
would
a
stress
vector
made
up
of
multiple
scales
integral
of
multiple
scales
just
know
about
the
local?
So
that's
the
evidence
for
why
we
would
do
that.
Then
we
got
into
the
discussion.
Well,
what
would
we
do?
How
would
we
actually
compute
the
momentum,
flux,
divergence,
okay,
well,
right
now,
the
plan
is
based
mostly
on
bob
hallberg.
Thank
you
bob
that
then
he
says
that
the
incur
current
implicit
shear
oriented
with
some
possible
wave,
enhanced
viscosities
that
you've
just
heard
about
will
stabilize
the
solution.
F
So
that
will
give
you
you
and
v
at
various
depths
k
at
n,
plus
one
or
an
update
at
u
points
and
v
points
and
then.
F
Is
that
you
can
put
an
explicit
wave
enhanced
inducement
induced
enhancements
from
waves
at
the?
U
points
you
can
do
it
explicitly,
which
this
is
so
good.
I
mean
trying
to
figure
out
how
to
do.
This
implicitly
was
a
bit
of
a
nightmare,
and
it
is
just
almost
too
good
to
be
true
if
this
works
and
the
idea
is
to
at
least
try
just
put
these
increments
in
and
just
increment.
These
updates
that
you've
made
implicitly
and
put
these
explicit
wave
induced
enhancements.
F
So
that
means
your
mixing
scheme.
Whatever
it
happens
to
be,
as
we've
seen,
might
actually
give
you
waved
enhanced
viscosities
and
diffusivities,
but
then
it
could
I'm
just
a
little
too
sensitive
here.
Then
you
have
to
give
it
some
of
these
wave
induced
increments,
all
right,
and
so
that's
the
plan
at
the
moment,
and
so
then,
what
we
have
to
do
is
we
have
to
come
up
with
some
modif
priority
modifications
that
we
have
to
do
to
the
code
and
these
mo
these
mixing
schemes.
F
So,
if
we're
going
to
that's
what
we
need
to
do
today,
if
we
can
or
get
as
far
as
we
can,
try
find
some
high,
medium
and
low
modifications
and
perhaps
even
dedicate
you
know
saying
things
that
are
might
be
possible,
but
it's
just
too
premature
and
give
them
a
zero.
F
So
a
preliminary
stab
at
that
just
to
get
started
is
to
parametrize
the
interfacial
stress
as
a
magnitude
and
direction
or
components
so
that
you
can
you
don't
have
to
have
the
down
gradient
or
down
stress
and
shear
align,
so
your
components
might
be.
F
Now
this
is
what
you
want,
and
what
you
take
away
explicitly
is
similar
to
what
you,
what
the
code
will
put
in
implicitly
in
that
implicit
viscosity
time
step,
so
that
you
actually
get
what
so
almost
what
you
want,
but
you're
using
that
implicit
step
to
keep
stable
and
that's
the
hope,
okay,
and
I
really
like
to
see
that
why
to
do
this
and
why
it's
a
high
priority,
is
we
look
at
some
less
angles
from
the
wind?
This
is
okay.
This
is
a
plot
I
have
from
one
set.
F
So
the
shear
and
stress
in
this
particular
examples
are
well
over
90
degrees,
from
the
stress.
So
assuming
that
the
shear
and
the
stress
are
aligned
is
not
a
particularly
good
assumption,
and
given
that
we
spend
so
much
time
trying
to
get
the
surface
stress,
I.e
the
wind
stress
right.
Maybe
we
should
put
some
attention
into
actually
distributing
the
vertical
a
little
bit
better.
If
you
look
along
this
line,
if
that
is
basically
assuming
that
the
stress
in
window
line,
all
these
points
would
fall
along
here
and
that's
not
particularly
good
either.
F
But
clearly
the
stress
tends
to
stay
more
aligned
with
the
wind
than
the
shear
does,
at
least
in
this
one
example,
and
we
need
many
more
to
see
whether
to
what
to
do
about
this.
Hence
the
some
of
this
infrastructure
building
to
see
whether
any
of
this
is
even
important.
F
Another
priority
would
be
to
put
the
default
capability,
enhanced,
k
and
m
these
diffusivities
and
viscosity
independently.
F
That's
already
done-
or
at
least
I
don't
think,
they're
independent
yet,
but
you
can
enhance
the
viscosity
and
diffusivity,
but
that
is
kind
of
should
go
with
updating
the
shape
function,
at
least
in
our
kpp
version,
because
that
reduces
the
these
coefficients
in
the
lower
boundary
layer,
not
the
upper
boundary
layer,
and
so
you
they
should
probably
go
together
to
try
and
get
these
counter
effects
at
least
more
aligned
with
what
the
las
are
showing.
As
far
as
going
below
the
boundary
layer,
that's
pretty
premature.
F
Now,
though,
I
know
kelvin
richards
has
always
said:
there's
lots
of
shear
at
small
scale
down
there,
that's
where
this
would
come
in
and
there's
also
some
thinking
that
things
depend
on
time.
History.
Well,
that's
premature
too,
as
I
say,
we're
not
sure
that
once
these
models
get
these
different
equations,
whether
they'll
even
produce
anything
hugely
different
or
whether
it
matters
and
that's
the
first
thing
we
need
to
find
out.
There's
also,
which
might
be
a
low
priority,
in
my
view,
from
the
evidence
I.
G
F
Is
to
reformulate
the
hpl
calculation
to
make
it
go
deeper
if
you
have
the
right
flow
now.
Clearly
what
you've
seen
before
that
it's
been
reformulated
to
give
you
deeper
mixing
with
the
old
equations?
What
do
we
need
with
these
new
additional
terms
in
the
momentum
equations
and
that
evidence
that
I
have
is?
F
This
is
just
straight
or
current
version
using
les
meats,
I
mean
that's
the
key
and
you
can
see
relative
to
what
I
can
get
from
looking
at
the
turbulence
itself,
which
we
don't
have
in
in
mom,
six
it
it's
not
a
it's,
not
a
bad
fit,
there's
only
one
period
of
time.
It's
a
very
weird
part
of
the
diurnal
cycle,
the
night
after,
where
there's
just
no
buoyancy
flux
where
there's
there's
a
deviation,
but
basically
that's
why
this
is
pretty
low
priority.
F
It
is
low,
relatively
good
fit
whether
you
put
waves
in
or
you
actually
have
a
calm
case.
So
I
can't
quite
argue
that
that
should
be
even
a
medium
priority
thing
to
do.
Others
may
have
their
evidence
and
that's
what
we
need
to
see.
What
evidence
do
we
have
on
what
to
go
and
start
modifying?
So
that's,
that's
all.
I
have
to
say
until
I
hope
the
discussion
at
11.
B
Well,
thanks
for
inviting
me
to
speak
in
the
ocean
working
group,
I
think
I've
done
this,
maybe
once
before
so
I
wish
I
had
thrown
in
more
equations
to
fit
in
better.
I
want
to
start
by
thanking
my
co-authors,
especially
lettie
roach.
The
others
are
listed
here,
all
been
really
instrumental
in
this
work.
This
figure
really
is
my
motivation
to
start
with,
that
shows
the
sea
ice
edge
in
the
southern
ocean.
You
see
this
incredibly
beautiful
structure
of
eddies
and
a
gradient
in
the
I
would
say
quality
or
fuzziness.
But
that
really
is
so.
B
I
think,
fine
structure
that
we
are
seeing
as
fuzzy,
but
what
it
really
would
look
like,
or
flows
on
the
scale
of
a
meter
at
the
ice
edge,
and
then
these
kind
of
interesting
transitions
to
maybe
even
a
somewhat
abrupt
to
a
much
larger
flow
regime,
and
why
would
we
care
about
flows
and
well
sea
ice
modelers
would
care
because
they
influenced
strongly
the
I'll
go
back
to
this
perimeter
of
flows
and
the
ice
ocean,
lateral
interface
and
therefore
the
ice
albedo
feedback.
B
But
we
don't
have
a
lot
of
real
detailed
evidence
of
that.
But
this
figure,
I
think,
is
kind
of
cool
showing
when
we've
composited
waves
on
occasions
of
high
ice
extent,
events
so
five
day
kind
of
events
of
extreme
ice.
What
we
see
here
waves
tend
to
be
propagating
away
from
the
ice
as
well
as
having
lower
wave
energies.
B
So
if
you
think
about
the
opposite
of
this,
you
would
expect
when
the
sea
ice
retreats
to
have
waves
propagating
towards
the
ice
and
higher
wave
energy.
So
there
just
appears
to
be
some
empirical
evidence
for
a
relationship,
so
the
other
factors
may
be
for
you
or
that,
as
you
expect
waves
to
be
propagating
into
the
sea
ice
beyond
the
ice
edge,
there
should
be
some
effect
on
the
ocean
in
the
presence
of
sea
ice
that
currently
we're
not
capturing
in
the
version
of
wave
watch.
That's
in
cesm2
for
two
reasons:
one.
B
There
is
no
arctic
in
that
grid,
but
where
there
is
sea
ice
in
the
subpolar
oceans,
the
waves
are
assumed
not
to
propagate
into
the
sea
ice
and
there's
no
wave
energy
introduced
into
the
ocean
in
the
presence
of
ice.
So
no
way
when
wind
waves
created
in
the
presence
of
ice,
so
I'm
interested
interested
in
changing
that
with
these
motivations
for
many
reasons,
sea
ice
modeling,
as
well
as
ocean
modeling-
and
I
already
mentioned
smaller
flows,
really
really.
B
B
The
size
5
model
currently
treats
flows
as
they're
all
the
same
diameter
and
they're
300
meters
in
diameter,
and
this
basically
puts
us
in
a
very
weak
ice
albedo
feedback
regime
from
any
lateral
lateral
melt
is
very,
very
modest
in
that
model,
because
the
perimeter
is
very
small
when
reality
is
that
I've
shown
you
pictures
of
the
ice
edge,
but
also
analyses
of
those
kinds
of
images
have
shown
that
the
distribution
of
flows
by
number
here-
maximizes,
of
course,
as
you
get
smaller
and
smaller
they're,
just
some
semi
logarithmic
in
this
diagram
power
law
like
in
this
diagram,
and
that
basically
means
that
the
majority
of
flows
are
small
right
and
and
we're
treating
them
all
as
300
meters,
which
is
you
know
here
on
this
scale,
so
it
depends
on
where
you
are.
B
There
may
be
no
flows
that
are
even
300
meters
in
diameter,
so
we
started
to
model
a
flow
size
distribution
in
size
and
we
had
always
modeled
the
flow
sea
ice
thickness
distribution.
So
we
have
a
history
of
modeling
a
distribution
already,
and
this
is
that
thickness
distribution.
So
we
look
in
the
model.
B
We
treat
five
categories
of
thickness
and
an
open
water
category
as
well
in
a
probability
density
function
at
every
grid
cell,
so
it's
kind
of
a
subgrid
scale
parameterization,
but
with
a
method
for
allowing
redistribution
among
categories
when
there's
ice,
deformation
or
thickness
of
growth,
sorry
thickness
growth
and
melt,
but
now
we've
added
a
second
dimension
to
that
distribution.
B
So
here's
the
thickness
and
now
we've
got
a
horizontal
dimension
here
of
flow
size
and
so
we're
treating
it
as
a
joint
distribution
and,
of
course,
the
variation
in
flow
size
is
harder
to
see
because
it's
spanning
this
wide
wide
range
of
scales.
But
if
we
looked
in
detail,
we
would
see
that
it
is
thickness
dependent,
and
so
it
is
important
to
treat
it
as
a
joint
distribution.
B
So
we've
coated
this
up,
we've
been
doing
this
for
four
or
five
years
now.
I
just
want
to
briefly
explain
how
the
main
important
process
processes
for
this
distribution
flow
size,
distribution
that
has
two
dependent
variables.
The
size
and
thickness
is
a
transport
thermodynamic
effects.
Deformation
rigid,
rafting
effects
on
wave
fracture.
This
is
the
biggie
that
really
affects
flows
dramatically
in
size
by
fracturing
them
in
the
presence
of
waves
and
waves
are
incredibly
effective
at
fracturing
sea
ice
even
fairly
small
waves.
B
But
the
thing
that
actually
changes,
the
concentration
of
ice
is
just
lateral,
melt
and
lateral
growth.
So
the
feedbacks
on
ice
albedo
feedback
with
the
influence
on
isobutyl
feedback
comes
in
through
the
thermodynamics,
but
we
really
need
wave
fracture
to
create
the
small
flows
that
have
a
lot
of
lateral
mild.
B
B
So
this
is
something
we
implemented
in
wave
watch
three
based
on
a
theoretical
model
and
some
observational
empirical
fits
that
were
created
in
collaboration
with
mike
malon.
So,
just
briefly,
a
little
bit
of
our
results
to
show
you
that
this
looks
like
it
is
important.
We've
been
implementing
doing
experiments
and
a
series
of
publications
and
now
testing,
and
I'm
going
to
actually
present
this
kind
of
intermediate
level
where
we've
done
most
of
our
work,
and
that
is
with
just
sea
ice
and
ocean
waves.
Coupled
so
we
don't
have
an
active
atmosphere.
B
Instead,
we're
prescribing
atmospheric
analysis,
and
nor
do
I
have
the
ability
to
show
you
directly
how
it
influences
the
ocean,
but
I
can
show
you
how
we
think
it
might
we're
currently
testing
a
fully
coupled
model,
but
I
don't
have
any
results
from
that
to
share
just
to
give
you
the
drama.
First,
in
the
southern
ocean,
where
waves
are
surface
waves
are
our
greatest
near
the
ice
edge.
They
do
certainly
propagate
into
the
ice,
especially
in
summer
in
the
southern
ocean,
where
winds
are
still
very
strong
and
surface
waves
are
still
quite
large.
B
We
are
seeing
significant
wave
heights
in
the
range
of
one
to
two
meters
within
the
sea
ice
edge.
This
should
have
labeled.
This
is
a
the
15
sea
ice
concentration
contour.
B
So,
in
the
way,
marginal
ison,
the
dotted
line
is
85
percent
sea
ice
concentration.
We're
seeing
substantial
wave
energies,
of
course
this
depends
on
our
parameterization,
which
you
may
question.
The
concentration
of
sea
ice
is
so
much
higher
in
winter,
and
the
gradient
of
the
ice
edge
is
much
more
compact,
so
you
don't
see
as
great
a
region
of
large
significant
wave
heights
in
the
sea
is
not
just
because
the
marginalized
zone
is
smaller.
B
Remember
without
resolving
flows,
I
so
told
you
all
flows
were
300
meters
in
diameters,
which
essentially
looks
like
the
you
would
see
any
any
color
it'd
be
basically
black
on
this
scale
for
perimeter.
So
once
we
start
to
resolve
flows,
we
really
really
substantially
increase
the
flow
perimeter
and
in
turn,
the
amount
of
melt
that
occurs
because
of
ice
albedo
feedback
is
greater
and
that
ice
edge
tends
to
retreat
in
the
arctic
and
just
overall
in
the
antarctic,
you
see
because
the
sea
ice
concentration
is
already
so
low
in
summer.
B
B
Sorry
moorings
here
that
I'm
going
to
compare
with
and
just
looking
at
the
distribution
of
wave
heights
here
as
a
probability
density
in
the
moorings
in
blue
and
the
model
in
red.
We
have
much
greater,
of
course,
number
of
observations
taken
in
the
model,
so
the
we
have
high
confidence
in
this
distribution
in
the
model.
What
we
see
is
that
the
model
tends
to
exaggerate
the
significant
wave
heights
compared
to
the
moorings,
which
is
also
confirmed
by
another
analysis
of
the
non-dimensional
wave
energy
versus
the
peak
frequency.
B
When
we
look
at
the
spectrum
entering
into
the
waves
into
the
sea
ice,
we
see
the
model
tends
to
exaggerate
swell
waves
and
under
emphasize
the
shorter
frequency.
Sorry
higher
frequency,
shorter
wavelength,
wind
waves,
so
we
think
that
there's
a
deficit
of
wind
energy
being
introduced
in
the
presence
of
sea
ice.
Due
to
that
scaling,
I
mentioned
of
just
decreasing
the
introduction
of
energy
by
the
open
water
fraction.
So
with
that,
I
think
I
can't
really
tell
what
time
I'm
at
maybe
a
little
bit
over.
L
L
All
right
does
that
show
up
full
screen
for
you
guys,
yep
good,
so
I'm
melissa,
moulton
kind
of
new
to
this
community
so
to
introduce
myself,
I
sit
in
the
ocean
section
at
incar
and
I'm
also
affiliated
with
the
university
of
washington
through
both
the
applied
physics
lab
and
the
civil
and
environmental
engineering
department
and
my
background's
in
coastal
physical
oceanography.
L
But
at
you
know,
really
small
scales
like
in
in
this
picture.
You
know
these
are
sensors.
I
deployed-
and
I
use
observations
like
this
and
and
models
to
look
at
processes
and
hazards
in
this
zone.
Things
like
rip
currents
that
people
would
encounter
on
a
sunny
day
at
the
beach
and
then
also
things
like
in
this
picture
during
extreme
events,
wave
breaking
resulting
currents
and
erosion.
L
So
I'm
I'm
talking
here,
I
think,
is
a
little
bit
of
an
outsider.
To
give
a
sense
of
you
know,
one
of
our
topics
that
that
frank
suggested
was
brainstorming
about
research
opportunities
that
these
new
wave
coupling
capabilities
may
enable.
So
one
area
is
the
coastal
region.
So
you
know
I'll
talk
a
little
bit
and
I
think
everyone
will
agree
that
there's
a
ton
of
interesting
topics
at
this
intersection
of
climate
and
coasts
and
and
human
impacts.
L
But
it's
it's
hard
to
bridge
these
scales
and
resolution
and
what
processes
are
actually
represented,
really
limit.
What
can
be
done
and
so
I'll
give
a
few
thoughts
to
start
discussion
on.
You
know
what
we
can
do
and
how,
in
terms
of
using
this
global
information,
to
look
at
impacts
on
on
much
smaller
scales.
L
So
first
in
terms
of
of
scales,
you
know
it
can
be
helpful
to
point
out
that
there
really
is
a
huge
range
of
scales
involved
with
with
coastal
processes,
and
I
I've
tended
to
work
kind
of
in
in
this
this
orange
zone.
And
it's
it's
really
interesting
to
think
about
how
we
can
bridge
this
really
wide
range
of
scales,
including
you
know,
spatial
scales
that
aren't
even
shown
here
in
terms
of
global
estimates.
L
And
then
these
long
time
scales
over
which
you
know
humans
are
really
influencing
both
climate
and
and
our
coastlines,
so
to
to
start
out
just
motivating.
Why
we
study
these
coastal
regions
and
also
to
point
out
why?
Sometimes
we
do
really
need
to
resolve
these
small-scale
processes
in
the
coastal
region.
I
just
wanted
to
bring
up
one
one
thing:
I'm
working
on
at
ncar,
which
is
trying
to
understand
coastal
flooding
resulting
from
tropical
cyclones
and,
in
particular,
high
precipitation.
L
Events
where
you
have
combined
effects
of
precipitation
and
and
storm
surge
and
waves
are,
are
a
pretty
important
part
of
of
storm
surge
and
a
factor
that
that
I'm
thinking
about
is
not
only
you
know
the
flooding
of
homes,
but
also
the
flooding
of
infrastructure,
and
how
that
can
release
contaminants
to
the
environment
and
have
big
impacts
on
human
health
and
and
coastal
ecosystems.
L
You
actually
have
to
go
down
to
meter
scales
to
get
to
estimate
this
flooding
with
with
reasonable
accuracy
and
another
difficulty
there,
given
how
high
resolution
those
simulations
need
to
be
is
that
we
often
need
to
use
ensemble
approaches
to
really
to
get
at
uncertainty,
for
example,
associated
with
different.
You
know:
sea
level,
scenarios
or
wave
climate
scenarios,
but
including
waves
is
pretty
expensive,
as
many
of
you
know
and
have
mentioned,
waves
contribute
a
smaller
fraction.
L
So
I
show
pictures
like
this
to
point
out
that
you
know
point
out
the
obvious
that
with
csm
we're
not
attempting
to
reproduce
this
kind
of
work
in
the
coastal
region,
and
I
think
the
question
is
you
know
what
what
can
we
do
with
waves
in
in
casm
to
inform
these
kinds
of
important
topics?
L
So
what
you
know?
What
can
we
do
well
at
a
global
scale,
and
so
I'm
not
an
expert
on
this,
but
from
some
digging
around
I.
I
was
surprised
that
you
know
these
global
scale
wave
simulations
actually
can
reproduce
wave
heights
and
at
coastal
moorings
with
reasonable,
reasonable
accuracy.
L
L
Squared
with
you
know
the
wave
energy,
but
still,
I
think
this
kind
of
information
can
be
good
enough
to
address
questions
like
how
much
energy
reaches
the
shelf,
for
example,
under
different
emission
scenarios
and
that
can
can
include
kind
of
integrating
over
a
number
of
you
know,
storm
events
and
and
other
things.
So
that's
that's
good.
I
think
that's
a
you
know
a
clear
example
of
how
how
even
at
the
global
scale,
we
can
say
something
about
waves
that
are
incident
on
a
coastline,
but
you
know
these
locations.
L
I
would
still
call
offshore
as
a
coastal
oceanographer
and
a
whole
lot
happens
between
those
offshore
moorings
and
and
the
coast
where
these
actual
impacts
happen.
You
know
where
waves
interact
with
infrastructure
or
people
that
are
that
are
at
the
coast,
and
so
here's
kind
of
a
quick
schematic
of
one
way.
I
think
about
this,
and
this
isn't
meant
to
be
to
scale.
But
you
know
these
red
dots
could
represent
wave
information
and
other
information
from
you
know
either.
L
Maybe
a
you
know:
30
kilometer,
global
climate
scale
simulation
and
previous
slides
showed.
You
can
actually
get
pretty
good
information
about
how
the
wave
climate
and
other
factors
might
be
changing
here,
but
there
is
quite
a
bit
that
happens
between
there
and
the
coastline
that
influences
what
the
actual
impacts
are.
L
So
I
tend
to
think
of
you
know
that
we
need
work
to
think
about
how
to
effectively
down
scale
this
information
to
the
impacts
via
inclusion
of
things
like
you
know,
interaction
with
bathymetry,
ice
vegetation
and
other
factors
and
then
also
characteristics
of
the
coastline
itself
that
affect
whether
flooding
occurs
and
erosion
and
topics
like
that,
and
then
I
also
wanted
to
point
out
that
you
know
there
are
places
that
maybe
aren't
included
yet
where
these
coastal
impacts
can
feed
back
to
to
climate.
L
You
know
some
have
been
mentioned,
tangentially
things
like
you
know
if
you
have
erosion
of
a
permafrost
coast
or
melting
of
landfast
ice,
or
things
like
that.
So
briefly,
I
just
wanted
to
show
a
couple
examples
of
where
you
really
need
to
consider.
What's
going
on
in
this
region
between
where
the
global
scales
are
global,
simulations
are
resolving
this
information
and
where
the
actual
impacts
occur.
L
One
that
I
wanted
to
bring
up
is
you
know
more
of
a
polar
topic,
but
since
that's
the
place
where
waves
are
wave,
climate
is
changing
most
dramatically.
I
think
that's
the
place
where
there's
a
big
opportunity
for
new
coastal
collaborations
here
and
I
won't
go
into
detail,
but
this
picture
shows
landfast
ice
on
coast
in
in
northern
alaska,
and
the
point
is
that
you
know
these
global
scale.
L
Wave
models
can
tell
us
something
about
the
wave
energy
out
here
and
it
actually
tells
us
something
pretty
decent
about
the
wave
energy
pretty
close
to
the
coast,
except
you
know
in
places
where
these
moorings
are
really
deep
in
this
in
this
ice,
where
the
coast
is
then
completely
protected.
L
So
this
is
a
case
where
you
know
taking
this
offshore
information
alone
would
not
tell
you
which
part
of
this
coastline
was
going
to
erode
other
other
examples
of
how
we
can
maybe
downscale
information,
I
think,
include
a
lot
of
very
reduced
complexity
models
to
be
able
to
do
the
kinds
of
ensemble
type
approaches.
We
can't
simulate
in
detail
things
like
coastal
morphologic
change,
but
there's
been
lots
of
work
done
on
how
to
parameterize
those
effects,
and
that
tends
to
be.
You
know
using
kind
of
simple
geometries
and
things
like
that.
L
But,
for
example,
you
know
you
could
take
wave
information
from
and
see
level
information
and
then
relate
that
to
expected
shoreline
changes
on
a
global
scale.
There's
many
other
examples,
including
you
know:
coastal
flooding,
arctic
coastal
erosion
topics,
barrier,
island
migration-
and
you
know
a
common
theme-
is
that
these
are
really
quite
reduced
complexity
models
in
order
to
be
able
to
address
changes
on
long
time,
scales
and
across
different
regions.
L
The
last
topic
I
wanted
to
mention
as
a
discussion
point,
is
you
know?
To
what
extent
can
you
know
I
mentioned
kind
of
these
reduced
complexity
models
and
using
our
csm
output
as
kind
of
a
boundary
condition
on
coastal
models.
But
to
what
extent
can
we
more
directly
use
regional
configurations
or
even
global
simulations
at
ultra
high
resolution
like
ihesp
or
something
you
know
that
include
waves
to
address
actual
coastal
topics?
L
And
so
you
know
these
capabilities
are
really
exciting
and
you
know
there
can
be
power
in
using
doing
this
nesting
in
the
same
code
base.
L
You
know
as
global
models
where
you
can
look
at
influence
of
things
like
enso
or
other
things
on
on
these
patterns
and
then
there's
added
power
if
we
could
couple
with
waves,
because
we
could
then
include
effects
of
stokes
drift
and
the
associated
eulerian
return
currents
on
on
trajectories
still,
I
would
argue
that
these
models
are
not
resolving
well,
the
very
important
factor
in
terms
of
these
connectivity
estimates,
which
is
you
know
how
the
larvae
actually
get
to
and
from
the
beach,
and
so
I
think,
even
in
these,
you
know
nice
super
high
resolution
configurations,
there's
always
going
to
be
a
point
at
which
we're
not
resolving.
L
You
know
some
of
these
smaller
scales
of
coastal
processes,
which
may
may
or
may
not
be
important
for
the
the
topic
of
interest,
and
so
some
of
the
things
I'm
working
on
are
ways
of
using
other
modeling
tools
to
parametrize
those
impacts
better
in
in
studies
like
the
one
that
I
just
showed.
L
So
you
know
that
that
was
it
and
I
you
know,
I
think
it's
it's
interesting
to
brainstorm,
how
these
things
can
use
be
used,
but
I
thought
that
it'll
be
really
interesting
to
just
see
when
these
products
are
out
there.
How
coastal
oceanographers
pick
up
these
tools?
What
what
aspects
they
trust
for
their
problems
and
and
where
that
goes.
A
Okay,
great
thanks
all
right.
We're
gonna
find
we're
gonna
wrap
up.
The
presentations
with
a
short
presentation
from
albert
is
showing
us
where
we
stand
today
with
and
in
the
csm
framework,
with
mom
6
and
wave
watch,
3
and
timings,
and
so
on.
So
helper.
A
C
A
E
E
C
Like
he's
in
here,
but
he
hasn't
joined
in
the
audio,
yet
I'm
not
sure
if
that
was
from
earlier.
If
we
just
joined.
I
Awesome
so
sorry,
for
that,
I
have
this
single
slide
summarizing,
where
we
are
in
terms
of
coupling
mum
six
and
we
watch
three
and
cesm.
I
So
we
just
recently
added
this
new
way
watch
tv
version
into
csm.
This
is
version
6.07,
and
so
this
new
version
currently
coexists
with
the
old
legacy
wave
hd
that
we
have.
The
new
is
the
new
one
is
optional,
but
it
will
replace
the
legacy
version
that
we
have
once
we
finalize
the
development
version
and
the
new
version,
this
new
incrementation
brings
in
many
bug
fixes
and
software
informants,
and
it
also
has
these
routines
for
the
partitions
partitioned
stocks.
Drift
computation
that
brandon
described
earlier,
and
also
it
has
these
ice
vape.
I
Coupling
changes
from
helen
letty
and
cece
cece
described
these
coupling
changes,
and
so
for.
As
for
this
new
version,
we've
coupled
this
with
both
pop
two
and
month
six
and
validated
the
pop
two
coupling
by
running
a
jra
simulation
and
we
pretty
much
get
the
same
boundary
layer.
That's
when
we
compare
the
old
way
watch
three
versus
the
new
wave
watch,
three
coupled
with
pop
two
and
we've
added
20
grits,
the
one
degree
upgrade
and
the
0.6
degree
trifolder
mum
6
grid.
I
I
This
is
brandon's
approach,
and
one
thing
I
want
to
highlight
here
is
that
we
are
using
this
new
neuopsy
feature
where
it
allows
us
to
export
or
import
multi-dimensional
fields
that
can
have
ungridded
dimensions
and
in
the
case
of
this
n
monochromatic
waves,
the
ungridded
dimension
that
we
have
is
six
or
n
number
of
wave
components,
but
this
also
allows
us
to
even
import
and
export
the
full
spectra.
We
don't
have
to
define,
say
25
fields
for
each
spectral
element.
I
I
This
is
shingley's
work
for
papua
new
ashri
coupling,
where
we
pass
an
enhancement
factor
to
cv
mix
based
on
the
surface
layer
average
and
projected
language
number,
and
currently,
I'm
in
the
process
of
testing
and
validating
that,
and
the
reason
we
are
implementing
this
is
to
sort
of
have
a
test
bed
for
future
near-term
development
activities
that
we
will
be
discussing
next,
and
I
guess
with
that,
I
can
leave
the
floor.
E
A
Great
thanks
help
her
okay,
so
we've
gotten
through
our
discussion
we're
only
a
couple
minutes
behind
our
intended
schedule.
We
have
plenty
of
time
left,
so
I
know
there's
been
a
lot
of
discussion
going
through
the
chat
so,
like
I
said,
I'm
hoping
that
the
panelists
will
kind
of
take
over
from
here.
I
did
in.
A
If
people
have
this
google
doc
open
put
in
a
couple
of
questions
to
maybe
guide
the
the
discussion
bill,
I
think
nicely
laid
out
a
potential
pat
development
pathway
with
some
indication
of
priorities,
I'd
like
to
see
some
discussion
or
consent
or
rejection
of
that
that
approach,
and
then
maybe
talk
about
you
know
where,
where
are
the
rate
limiting
steps?
What
are
the?
What
are
the
pieces
of
that?
A
That
we
have
to
wait
for
versus
what
things
can
proceed
in
parallel
in
terms
of
our
ocean
working
group
we've
basically
dedicated
half
of
alperi's
time
that
is
dedicated
the
ocean
working
group,
which
means
a
quarter
of
his
time
to
this
wave
project
for
the
next
say
six
months.
So
that's
a
resource.
A
We
have
in
hand
what
else
what
what
else
is
available
and
what
do
we
need,
what
observations
or
additional
less
calculations
might
we
need
to
help
guide
this
implementation
and
are
there
you
know
between
the
sea
ice
and
the
coupling,
the
boundary
layer,
physics,
etc?
Maybe
some
of
the
human
dimension
applications
are
there
any
conflicting
requirements
that
we
might
need
to
think
about
and
try
to
resolve
so
so
anyway,
I
will
turn
it
back
to
panel
list
and
I'll.
I
guess
I'll
start
calling
on
people
gokan.
You
had
your
hand
up.
A
C
I
have
a
question
for
baylor
the
sailor
and
it
may
be
too
technical
for
this
and
if
so
the
moderators
can
say
skip
it,
but
you
said
something
about
the
coupling
between
the
stokes
and
the
in
the
sort
of
orlarian
flow,
and
you
said
it
sounded
like
you
were
saying
there
is
no
coupling,
but
there
was
a
if
you
expand
the
product
out
in
the
kinetic
energy
term.
For
example,
you
get
a:
u
s!
Dot!
U
h!
Term!
G
Know,
if
you
remember
enough
to
know
yeah
yeah
another
there's
a
linear
coupling
and
so
a
linear
coupling
doesn't
jump
across
scales.
You
need
a
non-linear
coupling
to
to
have
a
scale
to
scale
like
to
make
your
parametrization
say
your
parametrized
wave
effects
interact
with
your
resolved
wave
effects.
You
need
to
jump
scales
to
do
that
and
this
equations
that
doesn't
have
that
capability,
which.
D
G
C
G
So
it's
a
u
times
a,
u
stokes
term
and
that's
that's
different,
because
they're
generated
from
different
parts
of
it
so
say
that
they
would
have
a
different
wave
number
in
general,
and
so
you
don't
get
that
sign.
K1
x,
plus
sign
k2
x,
flipping
out
of
from
scale
to
scale
the
way
you
do
in
turbulence.
So
it's
it's
a
simpler
problem
because
of
the
overriding
simplifications
in
the
wave
averaged
equation
set
in
the
beginning.
So
I
can.
F
G
C
Since
the
equations
went
by
quite
fast,
so
this
might
be
for
baylor
the
sailor
or
random
question.
So
in
the
momentum
equation,
every
single
term
is
modified,
excluding
the
diffusion
term,
so
the
diffusion
of
momentum
is
still
acting
on
the
eulerian
and
presumably
there's
a
reason
for
that.
C
So
that's
one
question:
the
tracer
equations
are
not
modified,
they
are
still
being
using,
but
they're
still
using
the
ordering
and
probably
there's
a
reason
for
that,
and
then
the
third
question
that
I
was
going
to
ask
is
given
that
the
momentum
is
now
most
of
it
is
using
the
eulerian
plus
stokes
when
we
compute
the
windstress
terms
or
other
surface
flux,
components
in
the
coupler.
What
relative
wind
are
we
supposed
to
use.
O
Yeah,
so
with
the
first
question
asking
if
the
friction
term
needs
to
be
on
the
eulerian
current
or
the
lagrangian
current,
I
think
that's
still
a
research
question
and
some
of
that
hints
at
what
bill
was
talking
about
related
to
the
directionality.
O
There
are
some
indications
that
mixing
is
actually
down
the
lagrangian
gradient,
though
you're
mixing
the
eulerian
momentum
very
much
a
research
question.
I
think.
C
G
So
it
doesn't,
actually,
it
probably
doesn't
make
too
much
of
a
difference
which
one
you
used,
but
the
surface
flux
is
importantly
different,
depending
on
how
you
treat
the
very
near
surface,.
E
Yeah
I
was
curious
about
when
baylor
was
talking
about
statistical
waves.
I
I
guess
the
the
question
is
most
of
the
time
the
ocean's
got
non-equilibrium
winds
and
waves,
especially
when
you
have
swell
things
like
that.
That's
one
part
of
the
the
recipe.
How
do
you
does
it
account
for
that?
The
other
thing
is
we've.
E
We
talked
quite
a
bit
about
wave
watch
three
and
I
guess
we've
kind
of
assumed
that
it's
just
perfect,
but
it's
got
some
pretty
empirical
terms
on
the
right
hand,
side,
especially
how
waves
grow
and
how
they
dissipate.
E
G
G
It
is
trained
against
the
level
of
non-equilibrium-ness
that
is
present
in
wave
watch,
which
seems
to
work
pretty
well,
but
it
is
also
not
capable
of
doing
a
lot
of
the
things
that
we
traditionally
would
think
of
as
being
important
for
non-equilibrium
waves
like
cross
and
swell
crossing
swell
is
missing,
but
it
is
not
missing
in
the
sense
that
it's
the
degree
to
which
it
reduces
the
magnitude
of
the
stokes
drift
on
average
is
present,
even
though
we
don't
actually
know
what
direction
the
stokes
drift
is
going
other
than
to
assume
it's
in
the
wind
direction.
G
So
for
the
specific,
very
rose-colored
lands
of
looking
only
for
langmuir
mixing,
it
seems
to
be
pretty
decent
at
doing
its
job.
But
the
questions
that
we
were
talking
about
in
the
chat
we
think
it's
going
to
do
a
quite
poor
job.
If
you
tried
to
predict
other
wave
statistics
and
that's
what
we
need
to
think
about
more,
but
it
might
be
good
enough
for
climate
use,
that's
that's
really
the
point,
and
then
why
did
we
use?
G
Wave
watch
was
essentially
because
we
didn't
want
to
become
wave
model
developers
and
noaa
was
happily
doing
that
on
their
own.
So
we
thought
it
was
better
to
partner
with
them
and
leave
that
to
them
and
then
just
highlight
the
pieces
of
the
story
that
we're
missing,
rather
than
rather
than
trying
to
do
the
fundamentals
that
the
wave
community's
been
working
on
for
since
hasselman
in
the
60s.
E
Oh,
no,
that's
fine!
It's
I!
I
guess
I
was
the
non-equilibrium
problem,
seemed
kind
of
well
from
a
research
point
of
view.
It's
pretty
interesting
because,
let's
say
the
winds
fall
and
the
waves
stick
around
they're
gonna
mix,
and
I
mean
it's
an
la
regime,
that's
really
small
and
that
that
happens.
That's.
G
It
it
surely
does
happen
in
the
in
the
real
world.
What
the
question
is
is
whether
it
happens
often
enough
that,
from
a
climate
application
perspective,
it's
worth
the
additional
cost
of
wave
watch
over
something
like
statistical
waves,
and
I
think
that's
the
system
that's
being
built
here-
is
really
up
to
the
user
to
decide
how
accurate
they
want
their
wave
field
to
be,
and
so
we
have
both
both
paths
available.
E
Yeah
I
just
mentioned
that
baylor
knows
already,
but
stephen
belcher
and
his
graduate
student.
They
looked
at
the
climatology
of
waves
of
of
essentially
wave
age
and
it's
pretty
non-equilibrium
over
most
of
the
planet.
G
E
The
other
thing
is
we'll
do
hurricanes.
That
kind
of
thing
I
mean,
if
you
make
hurricanes
in
csm,
I
guess,
but
with
super
complicated
wave
fields.
E
You
answered
the
ques,
you
put
your
good
thoughts
on
that
question,
whether
you
should
be
running
wave
watch
under
a
hurricane
or
the
applicability
of
etc.
O
Yeah
I
mean,
I
think,
a
hurricane
scenario
is
one
where
I
would
expect
the
the
theory
wave
stuff
to
start
to
do
a
bit
less
of
a
a
good
job.
You
know,
because
we
have
such
a
different
way
field
on
the
right
hand,
side
of
the
storm
versus
on
the
left-hand
side
of
the
storm
because
of
this
kind
of
coupling
with
the
waves
that
are
going
in
phase
with
the
storm
wave
watch
itself,
especially
the
latest
versions
of
wave
watch
three.
So
to
kind
of
go
back
to
your
first
question.
O
Wave
watch
has
a
number
of
source
term
packages,
so
a
number
of
these
parameterizations-
maybe
you
know
this,
and
so
these
would
be
open.
I
think
to
the
group,
if
you're
taking
the
latest
version
and
these
more
recent
versions
of
wave
watch,
definitely
have
better
dissipation
functions
that
are
meant
to
actually
end
the
wind
input
term,
especially
to
capture
the
hurricane
fields,
and
they
do
a
pretty
good
job.
E
O
A
Absolutely,
since
I
mean
the
motivation
for
this
discussion
theory
wave
data
wave
is
largely
driven
by
cost
questions.
I
thought
halper.
Could
you
share
that
last
slide?
You
had
on
timings,
which
seemed
not
as
scary
as
the
numbers
baylor
was
putting
up.
A
I
So
for
these
performance
results
I
just
turned
it
off,
but
yeah
for
validation.
I
was
turning
on
all
of
the
available
history
fields,
but
yeah
for
the
performance.
Again,
these
results
are
very
preliminary.
I
did
not
optimize
the
thumb,
step,
time
steps
or
the
processor
layout,
but
if
you
look
at
the
case
where
we
run
way
watch
three
with
the
course
grid,
the
cost
is
pretty
minimal,
like
two
percent
of
month
six,
but
if
you
run
them
on
the
same
grid
they
watch.
I
A
B
Numbers
are
well,
I
didn't
run
it
with
mom,
but
these
it's
kind
of
more,
my
impression
that
it's
comparable
to
the
ocean.
But
well,
I
guess
you
mentioned
opera
that
you
actually
turned
off
the
history
file,
which
I
think
is
probably
a
big
factor,
because.
I
B
A
Okay,
bill.
F
Yeah
I
wanted
to
just
float
a
couple
of
things
I
mean.
Is
it?
Is
it
a
plan,
or
could
it
be
part
of
the
plan
to
have
a
static
version?
6.07,
the
one
that's
just
been
implemented
cc.
It
looks
like
your
waist
ice
stuff's
been
put
in
there,
so
that
would
be
a
good
thing
to
at
least
discuss
whether
that's
static
and
and
also
we
seem
to
be
running
on
the
on
the
ocean
grid.
Whatever
it
is,
would
we
want
a
static
wave
watch
grid
that
we
ran
wave
watch
on
and
then
interpolated?
F
I
don't
know
the
answer
to
that,
but
I
think
if
we
get
some
guidance
here
at
least
for
this
first
implementation
and
beyond,
it
would
be
helpful
to
know
if
we
can
simplify
things
like
that.
B
Sure
I
have
no
idea,
but
just
as
far
as
the
resolution,
I
think
we
need
in
the
sea
ice.
We
haven't
really
tested
it
because
I
haven't
spent
any
time
changing
grids
which
is
part
of
our
future
plans.
But
my
guess
is
that
the
pop
grid,
you
know,
is
kind
of
optimal
for
the
sea
ice
in
some
regards
because
of
converging
meridians,
and
so
it's
only
increasing
the
latitude
separation,
which
is
already
fairly
decent
at
the
one
degree
model
grid.
G
E
A
E
F
So
I'll
ask
a
question:
back
to
brandon
I
mean
it
looks
like
we
should
plan
to
bring
in
not
only
the
six
surface,
stoke
strip
components
but
have
a
potential
for
a
variable
exponential
decay.
It
could
be
all
the
same
for
now,
but
it's
we
should
like
you
had
in
your
equations.
I
think
you
had
a
decay
sub
I,
so
we
could
just
assume
that
that's
that
could
be
different.
O
So
I
guess
the
the
point
bill
would
that
I
was
trying
to
make
there
is
that
we
could
have
six
different
decay
scales,
but
they
can
be
fixed
in
space
and
time
right,
exactly,
I
think,
yeah
I
think
that's
probably
sufficient,
although
you
could
of
course
consider
you
know,
passing
whatever
the
peak
wave
number
is
of
that
specific
partition
as
well.
O
If
you
think
those
details
are
are
worth
it,
I'm
not
sure,
once
we
get
up
to
n
equals
six
or
so,
if
that
will
make
much
of
a
difference.
F
F
We're
not
sure
whether
this
has
an
order,
one
impact
on
circulation
or
does
almost
nothing,
and
I
think
finding
that
out
will
dictate
an
awful
lot
of
how
much
more
effort
we
put
into
various
things,
and
we
also
want
to
make
sure
that
it's
easy
to
experiment
in
various
parameterizations
and
stuff
and
what
you
want
to
look
at
so
any
more
discussion
about
whether
that
one,
whether
the
output
from
a
one
cycle
ocean
ice
run,
would
be
sufficient
for
a
lot
of
people.
That
would
be
good
too.
D
Okay,
but
I
see
what
what
I
was
thinking
is:
I
see
potential
in
lots
of
different
directions
here
for
budget
chemistry
and
even
also
outside
of
the
biochemical
realm
for
satellite
simulation
of
surface
roughness
emulators
right
in
cesm
that
we
have
to
do
other
things.
F
Well,
if
you
could,
why
don't
you
try
and
see
if
there's
anybody
in
that
community
would
say
please
here's
the
wave
parameters
I
might
want
and
then
we
can
at
least
try
and
put
in
some.
You
know
dummy
kinds
of
transmissions
and
things
that
would
be
useful.
D
Yeah,
I
I
could
reach
out
to
rick
weininghoff,
who
is
really
keen
to
think
more
about
parameterizing
waves
and
gas
exchange
processes
and
see
what
he
has
to
say.
D
J
F
F
B
C
J
So
there
was
a
comment
a
couple
of
minutes
ago
about
holding
off
on
putting
in
the
effects
of
the
waves
on
gas
exchanges
and
things
until
we
understand
it
better
and
can
capture
the
observations
from
a
climate
modeler's
perspective.
I
think,
there's
an
awful
lot
of
value
in
exploring
whether
there
are
amplifiers
or
sensitivities
that
that
could
arise
from
that
coupling.
Even
if
you
can't
demonstrate
that
it
is
accurate
based
on
observations.
J
As
long
as
we
do
something
plausible,
you
could
write
a
lot
of
papers
to
help
explore
kind
of
the
the
fat
tales
of
our
distributions,
and
so
I
would
not
wait
until
we
have
kind
of
solid
observational
calibration
of
our
parametrizations
to
start
exploring
these
ideas.
I
think
I
think
they
are
very
much
worth
exploring
kind
of
in
a
more
speculative
way.
D
Yeah,
if
I
can,
I
want
to
follow
up
on
bob's
comments,
so
there
are
existing
parameterizations
for
c-state
dependent,
rc
fluxes,
although
not
maybe
not
very
accurate,
but
they
depends
on
common
wave
statistics
like
significant
wave
height
or
a
wave
stiffness.
D
I
D
So
some
wave
parameterizations
are
sensitive
to
how
well
you
you,
you
get
the
wave
statistics,
but
some
parametrization
is
not
so
because,
because
our
goal
is
not
to
like
approximate
the
wave
climate
right,
so
that's
the
the
the
the
goal
of
wave
watch.
But
so
our
goal
is
to
do
a
good
wave
parameterizations
in
csm.
So
we
really
need
to
consider
the
different
wave
parameterizations.
We
want
to
include
in
csm.
F
So
so
chin
can
I
ask
you,
I
think
we
do
need
somebody
to
at
least
talk
to
and
guide
us
in
in
any
of
these.
Even
the
preliminary
ones
like
bob
was
mentioning
just
so
we're
not
way
way
out
of
do
you
have
any
suggestions
on
who,
on
the
gas
thing
I
heard
I
heard
rick
ronikov
might
be
at
least
approachable,
and
perhaps
what
about
on
on
anything
else,
just
somebody
to
to
talk
to
and
get
some.
F
Because
some
of
it
gets
very
elaborate
very
quickly
and
we
need
to
have
a
sensible
approach.
I
think
for
this
first
implementation.
D
D
He
is
interested
in
looking
at
the
arc,
assisted
c-state
dependent,
rc
fluxes,
so
I'm
not
sure
whether
he
will
actually
implement
the
parameterization
in
csm
but
yeah.
I
don't
know,
I
don't
know
who
is
going
to
do
the
job.
E
E
He
knows
an
awful
lot
about
the
topic
thanks
peter
bill,
I
was
gonna.
Ask
you
if
I
could
just
switch
gears
just
a
tiny
bit
whether
you
were
thinking
of
all
the
changes
to
the
boundary
layer,
parameterization
that
you
worked
on
the
last
few
years,
whether
that's
part
of
the
intent
of
modifying
bomb
six
or
not,
or
what
is
your?
E
F
E
E
Yeah
brandon
mentions
that
he's
worked
a
lot
with
luke
at
princeton
stuff,
like
that.
I
was
curious
because,
when
waves
break,
they
essentially
create
a
surfing
effect
which
really
increases
stokes
drift
quite
a
bit
and
maybe
that's
too
much
of
a
researchy
kind
of
idea.
E
But
there's
potential
that
when
there's
braking
you
end
up
with
a
lot
more
wave
effect,
especially
as
it
goes
into
the
stokes
drift.
Then,
if
it's
not-
and
maybe
maybe
that's
just
a
research
question.
F
Yeah,
it's
certainly
a
wave
watch
wave
modeler
question.
I
suspect
they
would
have
to
try
and
come
up
with
such
a
enhancement.
I
would,
I
would
think,
but
right
now
you
know
we
have
zero
stokes
drift.
So
it's
not
too
hard
to
do
better
than
that.
F
Let's
take,
let's
take
sensible
steps.
E
No,
I
I
greatly
I
give
kudos
to
all
the
worker
bees
who
are
actually
trying
to
actually
take
fancy
ideas
and
turn
them
into
a
real
model.
It's
it's
really
good.
E
Oh,
no,
it's
gorgeous
it's.
If
you
look
at
the
breaking
statistic
as
a
function
of
wave
age,
it's
just
it's
just
fabulous.
A
A
A
J
So
this
is
not
weight
related
to
waves,
but
it
is
related
to
the
heat
budget.
I
think
bill
put
out
very
early
in
the
the
conversation
today.
The
idea
that,
if
the
atmospheric
people
are
giving
up
once
they
get
below
about
a
tenth
of
a
watt
per
meter
squared,
we
shouldn't
try
to
do
any
better
and-
and
I've
been
mulling
this
one
over
for
a
few
hours
and
it
just.
J
J
But
regardless
I
would
be
a
little
bit
reticent
just
to
take
atmospheric
numbers
where
radiative
damping
to
space
limits.
How
long
things
to
accum
can
accumulate
and
necessarily
apply
the
same
thinking
to
the
ocean
where
heat
in
the
deep
ocean
can
be
sequestered
for
and
accumulate
for
for
a
thousand
years.
J
I
I
I
think
we
probably
are
going
to
want
to
be
a
little
bit
more
con.
More
do
a
better
job
of
conserving
enthalpy
in
the
ocean
than
our
atmospheric
counterparts
are
willing
to
to
do.
F
So
bob
would
with
it,
but
when
frank
was
mentioning
this
global
fixture
to
do
that,
your
global
fixing
at
about
the
0.01
watt
per
square
meter
might
that
was
for
your
fear
that
it
might
accumulate
locally
somewhere.
But
I
think
we
have
to
be
careful
yeah
and
one
of
the
big
reasons
to
do
this
is
to
make
sure
you've
got
a
budget
that
sort
of
closes
so
that
if
somebody
puts
something
in
and
screws
it
up,
you
know
and
right
now
the
the
atmospheres
aren't
conserving
things.
J
So,
on
the
ocean
side
we
conserve
enthalpy
heat
separately
from
kinetic
energy
and
we
don't
convert
dissipated,
kinetic
energy
into
enthalpy
right.
The
enthalpy
budget
is
something
we
close
to
like
18
or
19
decimal
places,
although
there
are
things
like
the
radioactive
decay
of
potassium
and
seawater
and
heat
from
the
metabolism
of
fish
and
other
things
that
all
start
to
come
in
at
that
that
sort
of
hundredth
to
a
thousandth
of
a
watt
per
meter,
squared
sort
of
level
that
we
could
take
into
account.
J
E
C
G
The
so
peter's
las
is
also
based
on
the
wave
average
equations,
and
there
are
some
things
that
happen
near
the
sea
surface
that
are
phase
dependent,
so
some
observations
show
that
there's
mixing
that
occurs
in
like
only
the
crests
of
waves
and,
in
addition,
there's
white
capping,
which
is
explicitly
disallowed,
because
it's
a
steep
slope
of
n,
so
those
pieces
stay
out
of
the
wave
average
equations.
G
The
wave
average
equations
seem
to
be
quite
good
at
describing
things
like
langmuir
turbulence,
so
there
are
very
near
surface
effects
that
are
missing
and
you-
and
I
mean
peter-
should
comment
on
what
he
thinks.
The
solution
to
that
problem
is,
but
that's
a
well-known
part
of
it.
What
I
was
talking
about
is
just
in
the
analysis
framework,
especially
the
wave
average
equations
really
only
depend
on
stokes
drift
and
the
stokes
drift
statistics
are
very
large
scale
and
very
slowly
varying
that's
not
actually
true
of
waves
themselves.
G
There
are
small
scale
couplings,
particularly
in
the
surf
zone,
but
also
in
the
open
ocean
of
wave
current
interactions
that
occur
on
the
wavelength
of
the
waves
themselves
and
are
therefore
not
describable
with
stokes
drift.
At
all,
so
we're
doing
a
decent
job
of
bringing
in
a
simplified
wave
a
simplified
equation
set
in
appropriately.
But
we
know
that
they're
missing
pieces.
E
In
the
interest
of
time,
no,
no,
no
one
of
the
one
of
the
cool
tests
that
people
have
done
in
the
past
I'll
make
it
real
quick,
is
to
take
actually
a
wave
resolving
les
with
a
wavy
surface
at
the
top
okay
and
compute
it
as
best
you
can
and
then
compare
it
to
cl
asymptotics.
E
C
E
A
F
Yeah
well,
I
think
I
have
some
names
of
people
to
contact
to
see
if
we
can
do
some
more
simple
things,
especially
on
what
we
might
want
on
the
erc
flux
issue,
especially
if
nikki
can
take
on
contacting
rick
or
just
yeah.
I
I
think
I
I
know
the
contact
now
and
then
probably
we
should
broadcast
a
a
more
of
a
detailed
plan,
but
I
haven't
heard
anything
the
I
I've
asked
the
questions
I
needed
to
and
I
think
I
got
answers,
except
for.
F
F
And
then
we
that's
static
and
then
we
just
keep
working
on
that
and
worry
about.
Interpolating.
These
things
might
might
be
a,
I
guess,
a
static
thing.
That's
not
varying
underneath
us
when
we
we
do
these
experiments,
but
that's
I
haven't
heard
enough
debate
on
exactly
what
grid
to
use.
Yet
that's
what
I'd
like
to
know,
especially
near
the
coast
and
things
like
that,
should
you
telescope
near
the
coast
that
kind
of
stuff.
A
O
O
E
F
J
Brandon
emc
is
actively
working
on
fixing.
This
jessica
is,
is
working
on
it
very
actively.
So
we're
not
we're
not
looking
at
having
to
identify
new
resources
to
make
wave
watch
work
on
a
tribal
grid,
it's
something
that
will
be
happening
for
basically
for
weather
forecasting,
problems
and
things
like
that
that
are
kind
of
outside
of
the
resource
scope.
Of
of
this
conversation
right.
O
A
Okay!
Well
thanks
everyone.
I
hope
this.
I
think
it
was
successful.
I
hope
everyone
shares
that
you,
I
think,
as
several
people
have
said,
this
is
an
exciting
new
direction
for
the
ocean
working
group
and
it
seems
to
be
a
lot
of
interest
and
enthusiasm
across
the
community.
So
something
we're
going
to
push
on
in
response
to
that
enthusiasm
and
in
touch
and
hope
to
see
everyone
in
the
winter
working
group
meeting.
So
so
long.