►
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
B
All
right,
let's
get
started
it's
one
o'clock
and
I
didn't
leave
any
time
for
logistics,
so
welcome
to
the
climate,
variability
and
change
working
group
meeting.
So
we
have
a
lot
of
excellent
talks
lined
up
this
afternoon.
This
is
being
streamed,
live
on
youtube
and
also
the
chat
will
be
saved,
and
that
includes
any
private
chats.
You
might
make
they're
downloaded
as
well,
when
you
do
that.
So
for
the
speakers,
you
have
a
10
minute
slot,
leaving
two
minutes
for
questions.
D
Okay,
now,
let's
start
the
meeting.
As
I
said,
the
meeting
the
talk
will
be
ten
minutes
and
with
two
means
the
question.
D
I
will
give
you
a
a
means
warning
and
before
we
start,
I
would
like
also
take
this
opportunity
to
stand
giving
me
about
his
long
and
the
leadership
role
in
this
working
group
and
the
congregation
for
winning
the
csm
disenglish
achievement
award
and
also,
I
would
like
to
thank
the
championship
and
the
peer
collector
for
their
series
for
this
group
and
they
stepped
down
in
the
screen
and
welcome
server
and
larson
joined
the
cbc
working
group
co-chair
now.
D
Our
first
speaker
is
nicole,
is
my
her
okay
and
she
was
talking
about
enso
and
tv
nico
nicola.
Okay,
now
go
ahead.
E
E
E
I
guess
a
spatial
imprint
and
then
this
can
actually
change
the
sign
of
what
you
see.
So
the
aim
of
my
study
is
to
investigate
this
modulation.
In
five
large
ensembles
and
we're
using
large
ensembles,
because
we
have
lots
and
lots
of
events
that
we
can
composite
over
and
truly
tease
out
the
role
of
each
of
these
modes
of
variability.
E
E
The
observations
are
in
the
top
left
and
I'm
showing
the
ensemble
member
for
each
large
ensemble.
That
is
most
highly
correlated
with
observations.
E
E
Okay.
The
second
question
you
might
want
to
ask
is:
can
the
models
actually
represent
specific
decadal
variability
and
today
I'm
going
to
use
the
pdo
index
or
the
pacific
decadal
oscillation,
and
this
figure
is
very
similar
to
the
previous
one?
We
have
observations
in
the
top
left
and
the
highest
correlated
ensemble
member
for
each
model.
E
You
can
see
again.
The
overall
pattern
is
well
represented
in
all
models,
but
we
do
see
some
biases,
particularly
in
some
of
the
models,
including
csm
over
the
kuroshio
region,
and
we
do
see
this
tropical
pacific
cold
tongue
bias,
that's
in
the
inter-annual
and
so
signal
as
well
in
this
decadal
pattern.
E
Okay,
so
let's
jump
straight
into
some
results
and
try
and
answer
the
question:
is
I
and
so
our
teleconnections,
modulated
by
the
pdo
over
north
america,
in
the
figure
you
can
see
here?
Each
row
is
a
different
large
ensemble
with
csm
large
ensemble
in
the
top
row,
and
I
show
again
red
and
blue
for
temperature
and
green
and
brown
for
precipitation.
E
The
left
row
of
each
set
of
columns
is
el
ninos
when
the
pdo
is
in
a
positive
phase
and
the
right
row
is
when
el
nino,
when
the
pdo
is
in
the
negative
phase.
So
we
can
see
a
few
things
here.
First,
the
models
generally
agree
on
the
large
scale
patterns,
although
there
are
definitely
some
local
differences.
E
The
temperature
pattern
is
most
strongly
modulated
over
northern
north
america
by
the
pdo.
Precipitation
is
most
strongly
modulated
over
the
western
coastline
by
the
pdo,
and
we
do
in
some
models,
see
something
happening
off
the
eastern
coastline
as
well,
and
we
see
the
largest
precipitation.
Modulation
is
actually
generally
offshore.
E
E
The
third
row,
so
the
pink
and
green
is
the
sea
level
pressure
anomalies,
and
you
can
see
this.
Basically,
this
big
pink
blob
is
the
aleutian
low,
and
what
you
can
see
here
is
when
the
pdo
is
positive.
The
illusion
low
is
strengthened
and
moved
a
little
bit
to
the
west
as
compared
to
when
the
pdo
is
negative.
We
also
see
a
small
increase
in
the
high
pressure
over
the
american
continent
and
this
could
be
what's
causing
the
temperature
anomalies.
E
We
see
in
the
top
row
as
we're
bringing
more
warm
air
from
the
south,
where
it's
warmer
up
into
the
northern
latitudes,
and
this
kind
of
small
high
pressure
anomaly
over
the
continent
could
also
be
causing
less
cold
pressure.
Events
to
happen
over
the
northern
north
us
in
the
bottom
row
we
have
the
zonal,
mean
zonal,
wind
and
I'll
direct
your
attention
to
the
bottom
left
panel.
E
First
here
the
black
contours
are
the
mean
state
and
the
colors
are
the
anomalies
for
el
nino
events
when
the
pdo
is
neutral,
and
so
you
can
see
that
the
mean
state.
The
winds,
that
kind
of
of
the
jet
that
comes
onto
the
coastline
of
north
america
is
strengthened
and
shifted
south
when
you
have
an
el
nino
event.
E
E
E
E
E
So
how
might
this
impact
people
so
instead
of
showing
composites
here,
I'm
showing
regions
and
you're
looking
at
the
whole
distribution
of
all
of
the
events
within
the
ensemble
on
the
far
left,
I
show
alaska
for
temperature,
and
here
you
can
see
the
probability
density
function
for
blue
when
we
have
neutral
neutral,
thanks,
yellow
when
there's
el
nino
in
a
pdo,
neutral
green,
when
el
nino
is
co-occurring
with
a
pdo,
negative
and
red,
when
el
nino
is
occurring
with
a
pdo
positive-
and
you
can
see
here
that
compared
to
the
neutral,
neutral
distribution,
the
temperatures
over
alaska
are
a
little
bit
warmer
when
you
have
an
el
nino
and
then
that
effect
is
slightly
larger.
E
E
The
middle
column
shows
the
same
thing
for
the
colorado
river
basin,
and
this
is
precipitation
anomalies.
Now,
and
here
you
can
see
that
not
much
is
going
on-
there
is
potentially
a
small
effect
of
el
ninos
compared
to
the
neutral
phase,
but
the
pdo
doesn't
really
seem
to
be
having
much
of
an
effect.
E
And
finally,
if
you
look
at
the
sandy
san,
diego
california,
again
precipitation
on
the
far
right,
you
can
see
the
strong
effect
of
the
el
nino
compared
to
the
neutral.
So
the
neutral
is
blue
and
you
can
see
that
the
distribution
is
definitely
changed
by
an
el
nino
event.
But
you
can
also
see
that
if
there
is
an
effect
of
the
pdo,
it's
pretty
low
by
that
clustering
of
the
green
red
and
yellow
okay.
E
E
This
effect
is
really
also
just
a
simple
superposition
of
the
pdo
and
enso
and
the
models.
And
finally,
if
we
look
at
the
probability
density
functions,
they
show
a
lot
of
overlap,
and
so,
if
you
were
looking
at
an
individual
event
and
saying
hey,
we
know
the
pdo
is
in
this
phase
and
we
have
an
el
nino.
D
B
A
Yeah
thanks
technically
that
was
very
interesting.
I
was
just
curious.
Do
the
amplitudes
of
el
ninos
enter
into
this
at
all.
E
D
F
D
G
Okay,
thanks
asia.
Can
you
see
my
slide
and
my
cursor.
D
B
Screen,
I
don't
know
if
you
want
to
view
this
slideshow
or.
G
Okay,
so
hello,
everyone,
my
name,
is
lei
jong
and
I
am
a
research
associate
working
at
the
university
of
colorado.
Boulder
here
are
the
collaborators
of
this
work.
G
In
this
study
we
analyzed
the
impact
of
indian
ocean
dipole
on
el
nino,
satsang
situation,
so
iod
and
also
are
two
dominant
in
the
annual
kilometer
modes
in
the
tropical
indo-pacific
region.
Climatologically,
the
systemic
temperature
is
highest
in
the
indo-pacific,
warm
pool
region
which
drives
large-scale
winds
converging
towards
here,
but
during
el
nino,
the
pacific
rock
escalation
is
weakened,
accompanied
by
substance
anomalies
over
the
maritime
continent.
G
For
example,
a
study
by
anomala
in
edo
in
2003
found
that
when
the
force
and
atmospheric
model
with
central
and
western
pacific
warming,
they
find
easterly
wind
anomalies
over
the
tropical
indian
ocean.
So
this
result
suggests
that
el
nino
can
contribute
to
development
of
positive
in
the
ocean.
Dipole.
G
G
So
this
suggests
that
the
relationship
between
iod
and
enso
is
very
complicated
and
we
have
not
fully
understood
it
yet,
especially
why
some
positive
iodines
co-occur
with
la
nina
in
this
study,
we
focus
on
investigating
the
impact
of
iod
and
so,
and
hopefully
our
results
will
help
us
to
better
understand
why
events
such
as
1967
occur
so
to
achieve
our
research
goal,
we
analyze
the
observational
data,
including
er
sst
version
3b
and
the
era
20c
analysis.
G
Meanwhile,
we
leave
the
ocean
fully
coupled
to
the
atmosphere
elsewhere,
so
we
have
10
consumption
members
by
slightly
perturbing
the
initial
conditions
and
ensemble
average.
The
results
isolate
the
impact
of
the
indian
ocean's
reserve
temperature
anomalies
on
global
climate
variability,
using
which,
in
this
study,
we
analyzed
the
impact
of
iod,
and
so
so.
First,
we
selected
all
the
observed
positive
iod
events,
which
is
the
x-axis,
and
we
analyzed
the
dgf
mean
new
3.4
index
during
those
years
and
compare
observed
values
which
is
blue
with
the
new
3.4
in
the
model
which
is
red.
G
We
find
this
result
very
interesting
because,
although
the
positive
iod
tends
to
cause
or
to
contribute
to
el
nino,
because
these
red
bars
are
positive,
we
do
note
that
in
some
years
such
as
these
three
events
see
in
the
model,
there
are
la
nina
instead
of
el
nino.
In
other
words,
positive
iod
indeed,
can
cause
laminia.
We
find
this
result
very
intriguing,
and
we
want
to
understand
why
why
the
positive
value
effect
can
be
different
between
those
two
categories.
G
So
to
answer
this
question,
we
perform
the
composite
analysis
of
these
five
years
and
it
is
three
years
separately
and
here
is
a
result
in
the
top
panel.
This
shows
the
sst
anomaly
and
the
surface
wind
anomaly
during
positive
iod
that
causes
laminia
in
the
model
and
the
middle
panel
shows
positive
iod
that
causes
el
nino
and
the
bottom
panel
shows
the
differences
between
them.
G
We
find
that
the
largest
differences
for
the
positive
iod
are
located
as
the
one
pole,
for
instance,
in
during
this
category.
At
this
positive
iodide
event,
we
find
the
largest
warming
located
at
the
so-called
summer
clinic
ridge
region,
while
during
this
kind
of
positive
iot
event
largest
warming
is
located
at
the
northwestern
tropical
indian
ocean.
The
difference
between
the
two
is
even
clearer.
G
If
you
look
at
the
bottom
panel,
so
the
summer
cannot
reach
warning
drive
with
not
large
scale,
normally
wind
anomalies
together
with
easter
event,
anomalies
in
the
topics
due
to
the
anomalous
zonal,
ssd
gradient.
This
kind
of
positive
iod
drives
large-scale
anti-cyclonic
anomaly,
which
extends
to
the
southeast
indian
ocean
and
causes
another
event
anomaly
in
this
region,
because
the
background
wind
in
this
region
is
southerly.
This
wind
weakens
the
background
wind
and
caused
sst
warming
of
the
west
coast
of
australia.
G
In
one
of
our
recent
studies,
we
find
that
the
southeast,
indian
ocean
warming
and
the
central
pacific
coating
are
closely
connected
to
each
other
and
they
can
amplify
each
other
through
interface
interactions.
This
sounds
easy
in
the
ocean.
Warming
can
enhance
specific
treatments
and
cause
cooling.
G
Acidic
cooling
can
contribute
to
southeastern
industrial
warming
through
both
oceanic
and
atmospheric
teleconnections.
So
these
results
together
suggest
that,
during
this
kind
of
positive
iod,
these
summer
cannot
reach
warming.
It
can
cause
southeast
in
the
ocean
warming
which
then
through
interface,
interaction,
cause
pacific
cooling
and
therefore
causes
la
nina
in
the
pacemaker
experiment,
but
during
this
kind
of
positive
iot,
because
it
cannot
cause
southeastern
the
ocean
warming,
it
actually
favors
el
nino
to
further
prolo
hypothesis.
G
The
iod
can
cause
another
event
anomalies
here
in
this
region.
Again,
it
is
against
the
mean
state
winds
and
therefore
it
favors
sst1,
but
during
this
okay
thanks,
but
during
this
kind
of
positive
iot
event,
it
actually
causes
not
sensitive
anomalies,
which
enhance
the
background,
winds
and
the
favorite
ssd
coding.
Because
of
this
kind
of
different
response
in
the
sound
system
news
and
to
the
positive
value
deforesting
this
favors
el
nino-
and
it
is
oh
sorry,
this
fear
is
la
nina
and
this
favorite
el
nino.
G
D
B
G
Yeah,
so
in
the
psp
maker
experiment,
we
used
historic,
forcing
plus
rcp
8.5
for
1920
to
2005
and
2006
to
2019..
We
removed
the
linear
trend
in
this.
In
our
results,
we
also
tried
to
remove
the
regression
onto
the
global,
mean
surface
temperature
and
both
give
us
very
similar
result.
So
we
think
that
external
forcing
may
not
play
a
major
role
here.
A
Thanks,
ladies,
it
was
a
very
nice
study,
so
one
question
I
have
so
you
have
taken
a
very
western
pacific
centric
view
on
this.
Did
you
also
look
in
those
specific
years
when
you
had
when
you
had
the
la
nina
in
conjunction
with
the
iod
positive
iod?
Did
you
also
look
at
the
state
of
the
tropical
pacific
and
other
possible
extratropical
influencing
of
influences
on
the
pacific
itself.
G
Okay
yeah,
so
I
guess
the
question
is
because
we
include
a
part
of
the
western
pacific,
perhaps
in
the
model
this
can
influence
the
answer.
A
Also
forced
by
large-scale
influences,
and
so
the
you
know,
for
the
indian
ocean
to
affect
the
pacific,
the
pacific
needs
to
be
in
a
proper
state
and
maybe
like
marijuana
modes
or
other
extratropical
influences
need
to
be.
You
know
conducive
to
what
to
the
evolution
of
the
system,
so
I
was
wondering
if,
in
those
particular
years
you
also
looked
at
the
larger
scale
conditions.
G
Yeah,
that's
a
good
question
so
because
we
are
analyzing,
the
indian
ocean,
pacemaker
experiments
and
the
ensemble
average
average
results
are
mostly
due
to
the
indian
ocean
forcing,
but
I
agree
with
you
it
could
be.
It
is
also
possible
that
indian
ocean
may
affect
some
other
regions
and
then
affect
pacific.
G
In
other
words,
there
may
be
another
possibility
through
this
instead
of
just
the
southeastern
ocean,
but
we
haven't
looked
at
that
yet,
but
we
can
look
at
itself
in
the
future.
Thanks.
I
Can
you
see
my
screen
yeah
yep?
Thank
you
so
much.
My
name
is
shim
yu,
I'm
a
phd
student
at
colorado
state
university.
Today,
I'm
gonna
be
talking
about
the
work
that
I've
been
doing
recently,
along
with
my
advisor
dave,
thompson
and
and
a
lot
of
other
collaborators
at
csu
that
is,
and
also
with
casey
patricio.
So
I'm
looking
at
atmospheric
response
to
an
extra
tropical
ssd
variability
over
the
western
specific
using
a
set
of
simulations
running
the
csm.
I
So
I'm
going
to
start
with
providing
some
background
and
motivations
for
this
study.
It
is
clear
that
atmospheric
circulation
can
influence
the
extratropical
ssd
variability
through
the
surface.
Hipfloxies
and
surface
finish
stresses.
However,
the
atmospheric
response
to
extratropical
ssd
variability
is
relatively
solid.
I
The
linear
theories
predict
that
any
surface
heat
flux
anomalies
in
the
middle
attitude
can
be
easily
balanced
by
small
changes
in
low
level
circulation
due
to
relatively
strong
horizontal
temperature
gradient
in
the
middle
latitude
and
the
atmospheric
response
in
the
middle
latitude
is
even
more
complicated
because
of
the
strong
internal
variability
and
the
dependence
of
atmospheric
response
to
inject
position
and
non-linear
feedbacks
from
extra
tropical
eddies.
Thus,
theoretically,
the
atmospheric
response
to
extratropical
ssd
variability
is
much
more
difficult
to
isolate
and
quantify.
I
Thus,
this
pattern
is
consistent
with
the
atmospheric
forcing
of
the
sst,
and
if
you
look
at
the
plot
on
your
right,
there
are
low
ceiling
by
pressure
anomalies
that
extend
downstream
of
this
corrosion
extension
region,
and
this
is
associated
with
the
southward
cold
direction.
Just
this
pattern
is
consistent
with
the
atmospheric
response
to
cooling
of
the
sst.
I
So,
in
order
to
test
the
observed,
d-lag
relationships
between
the
extratropical
atmospheric
circulation
and
ssd
variability,
we
assessed
experiments
run
on
two
different
configurations
of
the
csm,
so
the
first
type
of
simulation
is
the
fully
coupled
aogcm,
which
the
atmosphere
mother,
is
copperton,
fully
active
ocean
with
sst
evolving
by
internal
physics
of
this
ocean
water.
So
here
we
can
assume
the
both
two-way
feedbacks
between
the
atmosphere
and
ocean
are
included
in
the
copper
simulation
and
the
second
one
is
a
prescribed
ssd
experiment
which
is
an
atmosphere
model
standalone
simulation
forced
by
data
ocean.
I
So
we
referred
to
cover
once
denoting
corrosion,
ssds
and
global
atmosphere
simulation
here.
The
simulation
is
forced
by
time-varying
historical
ssds
over
the
western
specific
and
also
with
annually
repeating
climatological
seasonal
cycle
over
all
other
regions.
So,
if
you
think
about
the
coupling
between
a
data
ocean
and
atmospheric
circulation,
you
you'd
expect
the
atmospheric
response
to
the
ssd,
forcing
but
not
the
atmospheric,
forcing
of
the
ssd,
because
ss
sst
fields
are
given
by
the
data
and
is
fixed.
I
So
this
shows
the
the
pattern
of
ssd
anomalies
that
used
to
force
the
koga
simulations.
So
you
can
see
the
largest
ssd
anomalies
are
focused
on
the
corrosion
reaction
extension
region
and
set
to
be
zero
elsewhere.
So
there
is
no
tropical
forcing
as
well
as
high
latitude
forcings.
I
I
So
you
can
kind
of
work
as
reference
and
the
left
column
shows
the
negative
leg,
which
is
the
atmosphere
which
shows
atmospheric
circulation
pattern
prior
to
the
peak
of
the
sst
normally,
and
the
right
column
shows
the
positive
leg,
which
is
the
case
for
sst,
is
leading
the
atmospheric
circulation
pair,
so
just
to
reveal
there
is
anticyclonic
anomalies
at
negative.
Leg
is
associated
with
the
third
world
war
temperature
direction,
and
this
pattern
is
consistent
with
the
atmospheric
forcing
of
the
sst.
I
And
if
you
look
at
the
upper
right,
there
are
low
silica
pressure
anomalies
that
lacks
the
peak
of
the
ssd
normally
by
a
month,
and
it
is
consistent
with
the
atmospheric
response
to
the
cooling
of
the
ssds.
I
Also,
the
lower
panels
in
the
red
box
are
the
result
from
the
copper
aogcn.
So
if
you
compare
the
upper
panels
and
on
the
lower
panels,
you
can
see
the
wizards
from
the
observations
and
the
coupled
argcms
are
actually
very
similar
to
each
other.
So
we
can
see
both
the
atmospheric
forcing
pattern
as
well
as
atmospheric
response.
Pattern,
are
well
well
represented
in
the
corporate
energy
same,
although
the
amplitude
of
these
low
pressure
anomalies
are
already
weaker
in
the
simulation
than
in
the
observation.
I
I
This
is
expected
because
the
set
ssts
are
given
as
data
and
thus
the
low
level
temperature
direction
by
atmosphere
is
incapable
of
influencing
the
ssd
field.
Rather,
the
result
from
kogaron
showed
this
low
pressure
anomalies
at
all
legs
and
by
construction.
Atmospheric
patterns
shown
in
the
kobash
runs,
are
or
representing
the
atmospheric
response
to
the
prescribed
ssd
forcing
so
this
closed.
I
Similarly,
between
the
atmospheric
response,
parana
koga's
run
and
the
inferred
atmosphere,
responses,
positive
lag
in
the
observation
suggests
that
the
lateral
pattern
is
actually
the
atmospheric
response
to
the
local
ssd
anomalies
over
the
western
specific
region.
I
I
So
if
you
look
at
the
left
two
columns
in
the
upper
panels,
the
real
result
from
the
observations
and
the
corporate
aog
gm
shows
that
the
at
negative
leg
there
is
a
warming
of
ocean
mixed
layer
through
the
downward
surface
photographs
of
fluxing
from.
I
Into
the
ocean
and
at
positive
lag,
there
is
a
cooling
of
sst
associated
with
the
upper
hip
fluxes
from
ocean
into
the
atmosphere,
so
at
negative
leg.
The
atmospheric
circulation
pattern
is
consistent
with
the
atmospheric,
forcing
of
sst
and
again
a
positive
leg.
Hemospheric
circulation
parent
represent
the
atmospheric
response
to
operative
hyphenoxis
and
if
you.
J
I
The
right
column,
the
clover
results,
show
the
overall
phalluses
at
all
legs.
Thus,
the
atmospheric
pattern
shown
in
koga
results
are
again
consistent
with
the
atmospheric
response
to
the
over
seed
fluxes
heat
fluxes
and,
to
summarize
the
observed
characteristics
of
atmosphere.
Ocean
coupling
in
the
pacific
are
fairly
well
captured
by
the
coupled
argcn
and
the
atmosphere
very
response
pattern
inferred
from
both
the
observations
and
the
coupled
air
gcm
is
recovered
in
the
prescribed
koga
experiment
and
the
response
pattern
is
highly
robust
in
the
observations
and
the
corporate
agcm
and
the
koba
experiment.
I
Oh,
I
I
can't
read
the
chat
story.
B
I
Yeah
yeah,
I
know
that
smear
no
paper
show
the
the
atmospheric
response
pattern,
especially
at
the
vertical
extent,
depending
on
the
relative
condition.
I
tested
this
with
one
degree
and
two
degree
standard
resolution
only
and
and
the
response
parameters
confined
to
the
lower
levels
like
below
700
half
pascal,
I
think,
is
consistently
the
low
resolution
version
of
the
similar
experiment,
but
yeah.
I
don't
think
I
can
do
the
high
resolution
experiment
again,
but
yeah
and
this
within
the
standard
resolution
it
didn't.
D
Okay,
thank
you
and
when
you
move
on
next,
I
will
give
the
team
fighting
this
stream
and
the
top
cooperative
connection
go
ahead.
Okay,.
K
Thank
you
so
the
first
I
like
to
thank
the
appreciate
the
contribution
from
my
collaborator.
I
put
the
name
there,
so
I
have
no
time
to
read
so,
and
you
can
see
this
is
a
teamwork
and
the
main
goal
is
very
straightforward.
Just
to
try
to
understand
this,
the
so-called
enhanced
jet
stream
waviness
feature
observed
over
the
past
40
40
years,
and
the
consensus
of
the
community
community
is
something
like
some
study
tried
to
argue.
K
This
is
the
enhanced
gesture
and
witness
in
mainly
in
northern
hemisphere,
it's
mainly
driven
by
this
arctic
amplification
or
co2
induced
aaa
effect.
The
idea
is
that
a
effect
is
warmer.
It's
faster
than
the
global
mean
temperature
that
can
reduce
the
nos
like
a
tropical
tropical
arctic
temperature
gradient,
so
that
can
reduce
this
that
enhances
the
women
and
reduce
the
jet.
But
here
we
try
to
argue
over
the
same
time
over
the
past
four
years.
K
Not
only
aaa
is
the
a
prominent,
but
at
the
same
time
there
is
the
cooling
phase
of
the
ip
ipo
interdicator
pacific
oscillation
stay
in
some
like
a
cooling
phase.
That
may
also
can
reduce
the
temperature
gradient
because
the
it's
a
type
of
water
scenario
like
like
article
warming,
can
reduce
the
temperature
building.
But
if
there's
internal
ipo
cooling
signal,
this
signal
can
also
reduce
temperature
gradient.
K
So
we
try
to
argue
probably
we
try
to
give
an
alternative
a
perspective
to
explain
the
same
thing
that
is
the
this
internal
or
we
call
a
surprise.
The
itcz
convection
during
summer
time
could
also
contribute
to
this
reduction
of
the
jet
stream
within
enhanced
waviness
but
reduced
the
strength
of
the
gesturing.
Something
like
that
and
we
use
a
model
and
some
like,
like
in-house,
like
experiment,
to
prove
this
and
show.
K
This
is
probably
another
way
to
explain
this
feature
and
from
the
in
the
right
hand,
side
you
can
see,
there's
a
lecture
six
of
panel
of
sorry
six
panel
and
the
so
this
six
channel
there's
a
six
seven
panel,
and
this
is
the
trend
of
the
jja
z200
from
er
e5
and
we're
using
this
one
to
show
you.
Actually,
this
willingness
is
enhanced.
K
Indeed,
over
the
past
four
years
along
the
jet
stream
northern
hemisphere
in
summertime,
you
can
see
there's
a
high
low,
high
price,
high
low
high
chain
of
a
high
low
high
pressure
anomaly
and
the
pattern,
and
if
you
remove
the
zonomine
component
and
this
with
whipped
cream
structures
become
more
clear.
And
so.
But
if
you
look
at
the
trend
of
the
mariana
wing,
you
can
find
there's
a
lot
of
enhanced
wavelengths
along
the
jet
stream,
and
this
is
a
zoning
component
of
the
z200.
K
So
you
can
see
the
maximum
height
rise
in
terms
of
dynamic
components
is
located
always
at
40
degree,
north
just
a
longer
jet
and
the
same
time.
Meanwhile,
if
you
look
at
the
trend
after
jj
sat
for
like
40
years,
you
can
find
this
global
warming
signal
very
clear,
but
also
there's
like
like
an
internal
cooling.
So
we
think
there's
a
probability
to
ipo
ipo
phase
shifting
to
this
cooling
phase,
and
this
is
a
kind
of
non-global
asset.
K
Just
using
this
trend
subtract
a
global
average
warming
signal,
so
you
can
see
there's
a
clear
cooling
phase
over
eastern
pacific
and
here
is
a
precipitation
just
like
a
jja
trend
over
the
same
period.
So
you
can
see
a
clear
is
draining
so
here's
the
shading
means
it's
a
draining
of
like
itcz
along
this
attitude
position
and
there's
another
warming
increase
of
precipitation,
not
warming,
the
increase
of
persecuting
muslim
or
is
the
maritime
time
continent
or
is
part.
K
So
we
think
that
this
structure
is
important
to
to
maybe
contribute
to
these
women
in
house
of
women.
So
then
we
first
thing
we
want
to
check.
Is
this
external
forcing
a
historical
run
from
different
sources
like
the
historical
csm1
large
example
for
remember,
add
them
together
same
layout
as
a
privacy,
one
and
seven
panel,
and
the
same
variable
same
same
as
non-global
and
zonal
things
and
the
75
historical
run,
certain
model
and
35
of
his
cm6
model.
K
We
put
it
all
together,
so
you
can
yeah
it's
really
busy,
but
you
can
clearly
see
that.
The
main
point
is
that
the
first
thing
is
a
tropical
snt.
Cooling
model
cannot
capture
and
just
a
model
if
the
raw
raw
trend
or
non-global,
or
give
like
an
el
nino-like
pattern.
No
matter
in
cinema
5
cm6
same
six
knots
astra,
but
a
csm
pretty
strong
and
the
rainfall
change
is
like
the
increase
like
the.
It
is
easy
to
have
in
our
vision.
We
see
this
a
decrease
of
at
least
a
decrease.
K
A
surprise
of
convection
along
rtcc,
and
another
thing
is
women
is
the
willingness
we
don't
see
very
clear.
Women
is,
if
you,
if
we
want
to
using
that
nonzonal,
d200
or
mariana,
when
to
quantify
that.
We
see
that
the
magnitude
is
much
smaller
than
observation,
and
the
observation
is
something
like
this:
the
union
here
is
a
per
decade.
So
this
lecture
one
meter
per
second
per
decade-
and
here
is
like
a
10.
K
But
if
you
look
to
multi
model
mean
she
sees,
the
zone
is
not
as
strong
75,
since
this
simplifies
improve
a
little
bit
and
capture
this
awareness
a
little.
But
it's
much
weaker.
It's
one
one,
one
third
of
the
observed
value
magnitude
and
with
marijuana
we're
not
that
strong,
then
the
cc
csm1
shows
some
wavelengths
here,
but
not
so
connected
to
throughout
this
longer
jet.
K
Another
thing
we
see
is
a
zonomin
component,
14
member
average
or
all
member
average.
It
seems
like
to
overestimate
a
tropical
warming,
because
you
see
the
orange
height
of
rice
is
much
stronger
than
observation
observations.
That
is,
the
the
red
one,
this
other
one
just
simply
copied
from
this
plot
from
this
curve.
This
line
so
idea
here
is
that
external.
K
So
another
thing
I
like
to
argue
is
I
see
the
article
warming
is
also
well
captured
so
by
rs
model
because
aaa
can
easily
trigger
by
this
the
co2
increase,
but
when
the
a
is
present
and
our
co2
forcing
is
a
strong,
but
this
in-house
weapon
is
not
that
clear.
K
So
we
think,
probably
due
to
this
the
the
a
little
bit
biased
the
simulation
of
a
tropical
convection,
because
in
observation
we
see
there's
a
cooling
signal
or
draining
itcz,
but
the
model
gives
us
c24s
and
gave
us
lecture
the
wetting
signal
there.
So
we
think
that
internal
topic-
asset
recording
just
like
ipo
related,
is
a
is
important,
but
we're
using
two
approach
to
prove
that
the
first
one
is
we're
just
using
model
itself.
We
don't
do
any
like
the
in-house
experiment,
so
we
just
simply
got
a
free
raw.
K
No
co2
falls
in
there
just
like
a
lung
controller
and
trimmed
to
very
short
40-year
spirit.
So
that
can
compare
with
observation.
So
we
have
many
many
like
a
sliding
window.
You
can
manage
like
like
40
years
of
period,
so
we
constrain.
So
we
ask
this
model
how
to
capture
his
women,
it's
by
using
like
a
spatial
correlation
to
constrain
that
one.
So
we
find
some
member
really
captured
his
waviness
and
some
pet.
Some
some
member
capture
like
opposite
face
like
the
blue
one,
and
then
we
composite
this
around
one
minus
a
blue
one.
K
So
we
want
to
see
because
we
are
constrained
only
over
along
the
jet,
like
a
20
knots
to
16
north.
So
we
want
to
see
that
what
would
happen
outside
of
is
that
with
latitude.
So
we
want
to
see
if
a
model
by
chance
capture
is
awareness,
because
we
ask
them
how
to
do
this
right.
We
force
model
to
capture
this
by
design
and
we
want
to
see
what
happened.
We
would
have
on
our
side
in
art
and
the
topics,
but
what
do
we
see
is
always
a
member
give
us
the
picture.
K
Is
that,
like
that?
Or
members
tend
to
have
this
cooling
in
in
the
tropics
like
like
ipo
and
the
joining
like
a
long
rtcc
and
no
article
warning
so
so
this
tells
us
something
like
in
this
model.
World
aaa
is
not
necessary
and
internal
cooling
becomes
more
important
to
to
to
concerned
with
this
a
in
health
awareness.
We
also
looked
at
foreign,
because
we
we
know
in
zampumin
doesn't
do
that
and
do
well
to
capture
enhanced
women's.
K
But
if
you
looked,
if
we
looked
at
individual
members,
some
members
still
to
do
better
than
other
members.
So
this
is
a
really
sorry,
oh
so
so
so
it's
the
same
thing
like
if
a
model
by
chance
capture,
internal
cooling
can
also
capture
wow,
it's
the
weakness,
so
we
want
to
so
this
last
two
plot
is
that
so
we
think
that
the
type
of
cooling
may
help
to
reduce
the
temperature
gradient.
So
this
observation,
this
is
the
model
and
it's
a
difference.
K
We
find
there's
a
cooling
right
and
we
think
there's
a
drying
is
a
precipitation
difference:
era:
minus
csm
large
example,
so
they
still
it's
a
cooling
may
help.
So
this
is
my
last
slide,
so
we
do
something
like
the
experiment.
We
just
put
this.
The
python
itc
is
suppression
and
this
enhanced
maritime,
the
precipitation
in
the
model
just
in
csm1
and
compare
control,
and
this
it's
just
like
a
snapshot
model
to
one
year
wrong
and
the
time
members
and
comparison
controller-
and
this
is
the
precipitation,
change
and
weaving
it.
K
So
we
find
the
model
can
capture
this
awareness
if
you
compare
its
observation,
not
completely
but
roughly
captured
like
in
house
of
enemies.
So
this
is
our
cartoon
summarize.
Our
main
idea
is
that
this
itcz
shift
due
to
the
ipo
cooling
and
generate
some
waviness,
and
this
maybe
there's
like
a
jet
stream
favors
the
tropical
instability
enhancement
and
there
is
the
wavelength.
So
there
are
two
things.
One
thing
is
the
cooling
favor
a
weakened
is
the
temperature
gradient,
but
same
time.
K
B
K
B
D
L
Oh
sorry,
yeah
hi
everyone
good
afternoon
today
my
talk
talkies
are
there
food.
L
L
The
reason
why
we
care
about
this
question
is
that
the
surface
wind
is
one
important
driving,
for
example,
ocean
modeling.
However
75
kilometer
scenario
results
show
that
there
is
a
increase,
a
robust,
increased
preservation
of
the
rtcd
region,
but
there's
no
much
change
in
the
surface
weighing
than
the
sea
level
pressure.
L
So,
however,
if
we
look
at
the,
if
we
look
at
the
opposite,
slow
pressure
signal
trend,
we
can
see
there's
a
decreased
low
pressure,
aluminium
over
the
tropical
atlantic
ocean
and
the
aim
pufa
amiibo
experimental,
also
suggests
decrease.
The
cellular
pressure
slow
play
decrease
the
slow
pressure
over
the
tropic
tropical
authentic
ocean.
L
Why
the
there's
not
much
changes
in
the
simplified
or
climate
model?
It
suggests
that
there's
a
very
nice
uncertainty
in
the
changes
of
the
surface
circulation
over
the
tropical
atlantic
ocean.
It
could
be
due
to
the
competing
effect,
such
as
the
thermodynamics
effect
over
warm
atmosphere
and
the
circulation
change,
such
as
the
heart
of
the
cell
expansion
or
the
changes
in
the
storm
track.
L
We
get
some
interesting
results
regarding
the
surface
circulation
of
the
tropical
of
atlantic
ocean.
That's
the
reason
why
I
give
this
talk.
The
apple
panel
shows
the
the
first
wolf
domain.
L
The
reason
why
I
choose
such
demand
is
to
count
the
ocean
ocean
dormant
that
is
shown
by
this
read
box
to
sell
the
computing
time.
I
try
to
minimize
my
model
grade,
so
the
source
boundary
of
my
model
domain
is
located
on
the
itcc
region.
L
The
results
prove
it
is
not
a
good
idea.
So
I
build
the
second
wolf
domain
to
cover
the
itcd
rating,
although
it's
not
phony
called
itc
day,
but
it
works
for
our
original
original
oceanic
modeling.
So
this
second
domain
to
the
first
domain
and
then
the
next
several
slides,
I
will
show
some
done.
Skinny
results.
L
First,
let's
look
at
the
precipitation
from
the
heart
of
the
center's
result,
model
result
and
the
work
done.
Skinny
result
in
the
historical
industries.
Three
rcp
scenarios
we
can
see.
Warf
did
a
fairly
good
job
of
the
extra
tropical
prestige
precipitation
zone
compared
with
the
hadley
center's
model
result.
L
However,
the
idc
world
is
unable
to
develop
let's
so
we
so
when
we
look
at
the
difference.
A
plot
between
rcp
manage
historical
scenario.
We
can
see.
L
There's
there
there's
no
matching
signal
of
the
appreciation
over
the
idc
regime
in
wolf
compared
to
very
robust,
increased
precision
over
the
itcd
region.
L
This
slide
shows
the
the
down
skinny
result
drawn
by
css
m4
rscp
2.5.
When
we
look
at
the
pre
precipitation
result
from
two
world
simulation
over
these
two
work.
Man,
we
can
say
it's
similar.
L
The
world
did
a
very
good
job
over
the
pres
in
the
presentation
of
the
extratropical
ranging,
but
the
idc
in
this
smaller
world
domain
is
unable
to
deal
an
output
and
the
idcd
in
the
second
world
domain
is
developed,
but
it's
where
it's
much
lateral
compared
with
the
css
compared
with
the
success
and
for
model
result,
then
we
look
at
the
difference
plot
between
rc
p,
minus
the
historical
scenario.
L
We
can
see
it's
a
very
similar
situation,
a
world
war.
Two
work
simulation
shows
similar
presentation
change
over
the
actual
tropical
region
compared
with
the
csm4
result,
but
over
the
adc
level
region.
There
are
the
increase.
The
precision
in
the
world
for
i2
result,
but
it
it's
it
is
it
is,
it
is
still
very
little
lateral
narrow,
runs
compared
to
the
csmo4
result.
L
Then
this
slide
shows
the
response
of
the
atmospheric
circulation.
The
first
three
columns
shows
down
skinny
result
driven
by
hardness,
centers
or
system
model.
The
first
column
was
just
done
skinny
without
drone
by
ccsm4.
L
The
upper
panel
starts
pre-preseration
change,
the
middle
and
the
bottom
panels,
just
geopolitical
high
chains
at
500,
500,
millibars,
200
millibars.
So,
first,
let's
focus
on
this
increased
precision
of
the
subtropical
western
atlantic
ocean.
Well,
you
can
see
this
this
this.
This
changes
in
the
presentation
are
comparable
among
these
three
scenarios.
L
However,
we
can
see
you
can
see.
This
is
a
kind
of
different
different
chains
in
the
in
the
500.
Your
potential
high.
Among
these
three
scenarios,
that's
because
the
increase
the
protection
alone
is.
It
is
consistent
with
the
legacy
or
geopotential
high
chains,
but
the
warm
dynamic,
thermodynamic
effect
of
the
atmosphere
will
come
tend
to
compensate
this
negative,
zero
potential
high.
L
As
a
result,
this
this
negative
steel
potential
high
chance
becomes
more
positive
with
the
rcps
the,
but
over
the
over
the
itcd
reaching
we
can
see
the
atmosphere
just
become
warmer.
The
geopolitical
potential
high
become
larger.
It
is
a
similar
similar
result
from
the
dark
skinned
result:
german
by
cssm4.
L
And
now
I
want
to
compare
the
two
simulation
of
the
world
over
the
two
different
domains
joined
by
cssm
form.
We
can
see
there's
quite
a
similar
precision
change
over
the
actual
tropical
region.
From
these
two
two
wolf
simulation.
L
B
L
L
Okay,
so
the
then
I
I
want
to
show
the
acetyl
pressure
and
surface
circulation
chance
to.
I
just
shorten
my
sort
of
my
talk,
so
we
can.
We
can
see,
there's
a
there's,
a
significant
cyclonic
acetyl
pressure
changed
just
over
the
subtropical
western
atlantic
ocean
that
is
corresponding
to
the
increased
presentation
there.
L
However,
we
also
looked
that
this
job
faster,
lactose
low
pressure
changed
in
this
four
worth
downstream
simulation,
but
there's
no
such
such
election
pressure
alarming
in
this
world
simulation
over
the
last
knight
domain
we
can
see.
This
is
a
very
big,
lengthy,
slow
pressure.
Alarming
is
from
the
the
intensification
of
this
sub
over
this
tropical
north
over
the
western
africa,
but
it
is
not
physical.
It
is
not
a
real
response
of
the
circulation
in
the
warming
climate.
L
I
don't
know
why,
but
it
must
related
with
the
domain
certain
boundary
domain
of
the
world.
So
I
just
took
my
summary:
the
take
home
take
home.
They
said
that
the
regional
model
domain
boundary
shouldn't
be
located
in
the
itcz
and
the
warmer
weak
surface
circulation.
Change
in
the
equatorial
authentic
ocean
could
be
a
result
of
the
competing
effect
of
warmer
atmosphere
and
the
increased
diabetic
hating
associated
with
the
convection.
L
L
Yes,
the
pessicon
is
the
the
result.
The
disambs
resulted
similar
with
the
results
warfare.
Without
all
this
large
domain.
B
L
F
All
right
thanks
mingo,
thank
you
all
right
and
just
a
reminder
we're
in
session
two
so
long
term,
climate,
variability
and
change,
and
so
I'm
your
chair
for
this
session,
sarah
larson
and
so
next
we
have
jana
anduan
evolution
of
long-term
temperature
trend
and
variability
in
cement,
6,
multi-model,
ensemble.
H
H
We
had
already
published
this
paper
in
geophysical
research,
letters,
the
motivation
and
the
objectives
or
of
this
study
is
to
conduct
a
robust
assessment
of
the
same
ip6
climate
model's
ability
to
capture
the
observant,
temperature,
train
and
reliability,
and
also
we
want
to
comment
on
the
single
large
ensemble
and
the
versus
multi
multi-modal,
larger
large
ensemble
stability
to
capture
the
observed
regional
temperature
train
and
reliability,
especially
in
the
sources
us
and
global
continent.
This
table
shows
the
models
we
use
in
this
study.
H
Until
we
close
this
paper,
we
already
had
30
33
or
normal
memory
on
the
models,
and
some
are
very
large
on
the
models,
and
some
are
very
small.
Only
has
only
has
one
only
has
one
other
member
in
this
table.
We
the
second
column
this
column,
is
a
temperature
trend
with
the
unit
of
sales
per
decade,
and
we
also
calculate
the
two
standard
deviation
of
the
arrow
to
test
the
significance
of
the
trend.
We
calculate
the
long-term
persistence
and
short-term
persistence.
I
will
introduce
it
later
after
that
we
want
to
compare.
H
The
data
methods
I
used
as
I
introduced
in
the
formal
slide
we
used
cmip6.
We
have
220
ensemble
members
from
the
23
climate
models
and
in
this
study
we
used
method
based
on
kumito
2030,
the
people
published
in
journal
of
climate,
and
we
do
some
further
study
to
discuss
more
about
the
7p6
properties.
H
The
long-term
trend,
persistence
calculation
method
is
non-parametric
trained.
Estimation
methods
the
significant
test
significant
test
calculation
is
calculation
methods,
are
long-term
persistence
and
short-term
persistence
as
introduced
in
formal
slide.
We
want
to
study
the
training
two
time
periods.
One
is
1901
to
2014.
Another
is
1951
to
2014,
because
the
as
the
following
time,
series
figures
show
so
within
several
several
decades
and
within
the
100
years,
100
years,
the
train
looks
very
different
compared
with
each
other,
especially
in
the
sources.
Us
I
will
introduce
more
in
the
following.
Slides
here
is
the
temperature.
H
It
corresponds
a
very
large
uncertainty
structure
in
the
same
ip6.
I
will
discuss
it
for
the
in
these
lines.
This
slides.
We
compare
the
temperature
increasing
trend
for
the
two
periods.
We
find
that
the
recent
several
decades
within
64
years,
the
temperature
trend
increased
much
greater
than
the
long-term
average,
almost
double
the
accelerated
rate
with
0.2
compared
with
0.08,
and
also
we
find
that
the
same
f6
model
can
capture
the
original
skill
reliability
much
better
for
the
short
term
period,
compared
with
a
very
long
term
average.
H
Then
we
calculate
the
trend
pattern
correlation
between
the
within
the
individual
model
and
within
the
220
ensemble
models,
the
same
ip6
and
also
the
model
observation
for
each
individual
model
to
get
some
idea
about
the
model
accuracy
globally.
They
are
the
global
content
results
globally.
The
special
pattern
of
the
trends
is
more
similar
among
the
or
normal
members
of
the
same
model.
H
It's
much
greater
than
the
intermodal
comparison.
The
intermodal
means
we
randomly
selected
the
allowable
members
from
the
220.
We
take
the
796
at
the
port
to
calculate
the
intermodal
pattern
correlation
and
this
one
is
a
modal
model
model
observation
pattern
correlation,
so
a
higher
number
represents
the
more
accuracy
we
have
for
the
random
selected
process.
We
repeat
the
process
for
500
times
for
500
times,
so
the
point
represents
the
500
times
average,
and
here
is
a
95
percentile
range
of
this
correlation.
H
So
we
want
to
compare
the
multimodal
light
on
the
most
mp6
and
also
some
single
model
with
very
large
on
normal
size.
We
want
to
compare
the
accuracy,
the
yellow
this
one
is
global
sources,
us
temperature
normally
in
average,
the
time
series
from
1901
to
2014
and
the
yellow
bar
shows
the
yellow
band
shows
the
the
failure
zone
where
the
climate
mode
cannot
capture
the
observation.
H
So
the
same
ip6
performance
is
much
better
than
the
single
model
light
on
them
due
to
the
large,
very
large
uncertain
range,
but
it
will
increase
another
problem
signal
to
noise
ratio
problems,
so
I
will
introduce
it
later
in
order
to
test
the
model.
Signal-To-Noise
paradox
problem,
because
the
model
sometimes
model
can
predict
the
observation
much
better,
but
it
is
caused
by
the
larger
something
the
model
itself.
So
we
want
to
test
the
balance
if
it
require
it
or
not.
H
The
rpc's
origin
of
predictable
component.
This
method
is
introducing
the
signal
to
noise
paradox
in
this
paper,
and
in
this
matrix
the
numerator
is
the
correlation
between
the
normal
mean
and
observation.
It
means
the
model
the
ability
to
play
to
capture
the
observation.
The
denominator
is
a
model
versus
model
correlation,
it
means
model
model
ability
to
predict
itself
in
the
perfect
situation.
The
rpc
should
be
one,
but
we
expect
value
to
be
smaller
than
one,
since
the
model
can
predict
itself
much
better
than
to
previous
observation.
H
In
this
study,
we
use
five
years
running
mean
time
to
calculate
rpc
and
to
in
order
to
test
the
significance
of
the
rpc.
We
use
the
sum
we
use
num
hypothesis
as
the
observed
temperature
trend
and
viability
line
within
the
sprite
of
the
model
ensemble.
It
means
if
we
replace
the
observation
with
any
random
slanting
model
itself.
The
rpc
should
be
the
same.
It
means
the
model
can
capture
the
observation.
H
So
we
want
this
situation
to
be
the
true
case
for
the
model
we
have
and
if
we
reject
null
hypothesis,
it
means
if,
if
we
replace
it
with
a
model,
the
rpc
should
be
different.
It
will
be,
we
will
repeat
the
random
select
process
1000
times,
so
we
will
show
the
95
percent
percentile
range
here
is
a
resource.
We
have
this.
One
is
a
semi
p6.
Another
is
normal
models
with
a
member
greater
than
10
greater
or
equal
equal
waste
10.
H
H
H
So
in
this
figure
the
rejection
rate
means
the
model
cannot
predict
the
observation,
much
better
than
printing
cells,
so
the
rejection
rate
always
occurring
in
the
yellow
range
yeah.
You
know
in
the
yellow
range.
There's,
no
wonder!
Well,
there's
no
wonder
some
ip6
has
a
very
low
rejection
rate
because
it
can
better
predict
the
observation
and
the
the
same
situation
occurring,
some
varying
in
some
relatively
small
single
model
like
the
ipsr
and
the
crm.
H
H
H
F
It
great
thanks:
do
we
have
any
questions.
F
Yeah
it
can
ask,
but
have
you
been
able
to
look
at
any
of
the
downscaled
versions
of
cmip5
that
are
kind
of
geared
towards
looking
at
the
more
specific
regional
differences.
H
Oh
no!
No!
No!
Yes!
It's
a
good
suggestion,
yeah
for
the
resolution
problem.
Now,
when
we
do
another
analysis
about
csm2
normal,
we
find
such
problem
with
a
different
resolution.
Results
will
be
slightly
different.
F
All
right,
we'll
move
on
to
the
next
talk,
and
so
next
we
have
ice
shoe
who,
with
variations
and
meridianal
overturning
circulation
and
their
influence
on
global.
F
D
M
D
This
work
is
in
collaboration
with
the
dear
meal
night
rose
and
bloom
maria
molina
and
the
rock
and
the
gary
strength
right
now.
The
paper
is
under
reveal
and
adrenal
fatigue.
D
The
atlantean
world,
overthinking,
circulation
or
fema
is
a
global
scale.
Ocean
circulation,
which
transports
upper
ocean,
water
and
salty
water
into
south
further
north
atlantic,
where
this
water,
loose
heat,
becomes
thinned
in
single
depths
and
flows,
southward
and
unwell
elsewhere
over
the
world
ocean
under
present
condition.
There's
no
deep
convention
in
the
cell
following
our
specifics,
but
studies
show
that
during
the
final
thing
about
2.5
to
5
million
years
ago,
there
might
be
an
ico
both
the
amos
and
the
precipitative
meridian
original
situation
for
chemo.
D
D
Two
is
how
the
famous
stability
will
influence
the
global
scale
motion
situation.
To
do
this,
we
use
the
csm1
at
one
degree,
horizontal
resolution
carried
out
five
experiments
for
the
first
four
extreme
experiments.
We
ran
for
800
years.
This
bread
wasn't
protein
on
for
500
years
and
earth
of
300
years.
For
the
last
experiment,
we
ran
the
simulation
for
300
350
years
with
spreading
water
for
an
hour
for
250
years
and
up
for
100
years.
D
This
firm
was
supposed
supposedly
compensated
elsewhere
in
the
world
ocean
or
the
global
head
experiment
and
compensated
by
the
non-supplemental
bristles.
Only
in
the
impact
experiment
for
the
last
one
we
put
adenosinol
into
the
south
pole
and
north
of
the
city,
and
this
soft
glass
is
completed
elsewhere
of
the
wood
ocean.
D
D
We
can
see
that
in
the
first
four
experiments,
both
these
simulations
show
a
class
with
a
mine,
but
when
the
first
water
thousand
common
completely
are
not
specific,
the
collapsing
is
slightly
slower
than
that.
The
condensated
by
the
global
ocean,
when
the
first
waterfall
stopped
at
the
year
500
the
experiment
responds
to
where
those
posing
is
recovered
immediately,
but
with
0.4
delayed
by
100
years,
but
they
both
overshoot
the
conference
room
and
then
recover
in
the
proceeding
site.
D
With
the
conversation
in
a
global
there's
no
p
mark
formation,
the
female
formation
is
only
happened
in
a
two
experiment,
with
the
combination
as
enough
specific,
that
means
the
collection
of
e-mark
itself
will
not
automatically
cause
a
set
up
a
female
in
our
simulation
so
to
test
a
sense,
our
demand,
whether
you
will
affect
the
amount
we
add,
the
adrenal
thought
in
the
suffering
not
presenting
to
both
the
set
of
edema,
and
then
we
look
at
atlantic.
We
do
see
a
weakening
of
the
amount
by
strengthening
of
the
tumor.
D
That
means,
if
we
force
able
to
set
up
it,
does
affect
the
circulation
in
the
atlantic.
This
is
the
marijuana
human
fountain
in
a
control
room
for
the
atlantic
and
the
pacific,
and
this
is
the
experiment
with
the
female
setup.
You
can
see
that
in
a
control
room,
there's
one
if
our
thing
in
the
atlantic,
but
posting
us
in
university,
when
the
team
set
up,
you
see
a
similar
meridian
stream
function
in
the
first
week,
I've
seen
the
language
in
a
controller,
but
right
now
it's
a
very
much
closely
flex.
D
D
When
the
amar
class
and
the
team
at
the
center,
we
see
the
strengthening
of
the
subcommittee
cell
in
the
south
atlantic
and
the
weak
meaning
of
that
in
the
north.
Seventy
in
the
processing
set
is
that,
with
the
setup
of
pima,
the
shallow
or
routine
circulation
in
the
pacific
is
awakening.
You
can
stop
for
today,
but
strengthening
not
personally.
D
We
can
see
that
weakening
of
the
amount
causing
awakening
of
the
merger,
the
varying
speed,
transforms
and
also
making
a
unique
through
flow
and
the
weakening
of
the
current,
but
the
strengthening
of
the
big
presence
and
stuff
with
the
setup
of
the
femur,
because
their
appearance
will
turn
about.
You
pull
the
wreaking
and
also
further
make
an
immediate
roof
load,
but
the
effect
on
the
advanced
current
is
small
and
the
sprint's
mean
of
big
heads
in
front
of
us
is
also
weaker
when
chemo
stays.
D
Up
next,
I
will
show
example
how
it
affects
sweden.
You
know
ocean
circulation,
I
see
the
changes
of
marijuana
flow
at
the
22
north
in
atlantic
and
its
pacific
figure
shows
the
atlantic
and
sides
at
20
years
in
the
control
run
and
in
the
file
experiment
in
control.
These
represent
nozzles
flowing
gulf
stream
represent
the
source
of
learning
return
current
of
the
amount
with
collapsing
amount.
In
these
four
experiments,
we
see
very
similar
changes
of
the
meridian
flow
in
the
atlantic,
awakening
of
the
dark
stream
and
almost
flat.
D
The
southward
depends
well
in
the
proceeding
salt
experiment.
We
also
say
this
awakening
of
the
dark,
beam
and
making
of
deep
kind
in
the
preceding
side,
with
the
inner
control
run,
we
see
it
is
not
worth
going
for
fuel
current
with
the
classical
amount.
Without
the
set
of
female
is
a
crucial
current.
The
deeper
heart
is
strengthened
and
surface
every
weekend.
D
D
So
in
our
experiment,
basically,
which
raised
our
four
different
states
of
image
and
humor
and
the
lcd
mark,
which
represents
the
modern
day
condition
and
in
iqa
max
and
in
actual
females
which
may
represent
the
future
climate
conditions
due
to
overwhelming
and
also
with
our
state,
is
that
we
have
active,
fema
but
grabs
the
iman.
This
situation
may
represent
a
homeless
event
during
last
degrees
of
freedom
and
last
one
is.
We
have
actual
amount
and
iq
fema.
D
D
D
D
F
Thanks
for
that
perfect
on
time
finish,
we
definitely
have
time
for
a
few
questions.
Yeah,
we
have
one
from
gokan.
A
D
D
Now
it
says
through
ocean
stratification,
but,
on
the
other
hand,
is
with
the
actual
ema
there's
a
flow
from
the
atlantic
into
the
pacific.
It's
about
a
field
struggle,
that's
where
the
collapse
very
much.
Death
flow
is
beginning
with
the
beacon
also
currently
to
the
strengthening
of
the
big
party,
but
it
is
small
that
we
can
still
see.
F
Yeah
I
had
one
you
showed
on
one
of
your
earlier
slides
that
two
of
your
experiments
had
a
amount
collapsed
and
then
they
oh
there
were
three
and
then
they
had
there
was
this
overshoot.
Do
you
have
any
idea?
What's
dictating
the
amplitude
of
that
overshoot,
because
it
seems
like
each
of
those
experiments
goes
to
the
same
close
to
you
know,
50
fair
drops.
D
D
Foreign
and
then
strengthening
will
pull
more
of
its
saltier
water
again
from
the
subtropical
atlantic
into
the
south,
polar
region
which
is
a
compound,
in
fact,
with
the
subsurface
loss.
Actually,
now
is
a
bit
of
pressure.
That's
why
it's
overshoot,
but
while
he
said
you
go
through
like
a
50s
ruler-
and
I
haven't
looked
at
that
yet.
F
N
Good
okay,
thank
you,
hello,
everyone!
This
is
shovely,
I'm
a
phd
student
from
the
university
of
california
riverside
and
today
I'm
going
to
talk
about
the
effects
of
historical
changes
in
sound
ocean,
hydaptic
and
storage
on
this
project.
I
collaborate
with
my
advisor
professor,
william,
and
I
also
collaborate
with
kaveli
andership
john
from
australia.
N
Well,
the
red
figure
here
shows
the
time
series
of
the
average
area
of
olympus,
so
we
can
find
that
due
to
human
activities
and
that
the
antarctic
oil
nuclear
has
increased
a
lot
in
the
past
few
decades
and
as
a
candle
of
greenhouse
gases.
So
the
genius
iron
concentrations
is
supposed
to
affect
the
climate,
particularly
in
the
southern
hemisphere.
So
in
this
project
we
wanted
to
know
what
is
the
rule
of
order
changes
in
southern
ocean
climate
change,
particularly,
we
focused
on
the
southern
ocean,
catapte
and
storage.
N
In
this
project,
we
used
to
use
one
large
and
thermal
simulation,
which
consists
of
42,
remember
members
with
historic
forces
and
is
named
as
historic
simulation,
and
meanwhile,
we
have
another
experiment
which
is
fixed-order
experiment,
which
consists
of
18
number
members
and
the
first
things
I
have
seen
the
students
one
large
and
limbo.
However,
the
older
concentrations,
including
both
stressful
golden
and
the
top
spirit,
codons,
are
fixed
at
1955
levels
and
is
named
as
historic,
fixed-order
simulations.
N
So
in
this
project
we
can
analyze
the
effects
of
historical
changes
during
1958
to
2005.
The
alarming
difference
between
the
two
simulations
here
and
here
are
some
of
the
results
we
started
with
another
engineers
in
the
atmosphere.
We
can
find
that
the
online
changes
leads
to
very
strong
power.
Intensification
of
the
surface
width
of
the
surface
when
it's
just
the
root
curve.
Impedance
here
and
the
other
induced
changes,
contribute
to
about
half
of
the
historic
surface.
N
Wind
stress
the
blue
curve
here
and
the
changes
of
the
surface
when
they
may
water
the
ocean
alter
the
winter
driven
ocean
circulation
in
the
southern
ocean,
which
is
a
decency.
So
we
found
that
the
alarming
merganometric
circulation
is
also
polar
shifted
and
also
enhanced
in
this
process.
Meanwhile,
we
found
that
the
anti-induced
smoke
is
also
enhanced
in
this
process,
which
partially
offsets
the
alarming
induced
smoke,
thus
producing
a
religion.
Mode
generally
follows
changes
in
the
alarming
mode
with
a
power
intensification
in
response
to
other
changes.
N
Besides
the
changes
in
ocean
circulations,
we
found
that
all
changes
can
also
alter
the
surface
hit
flex,
and
the
panel
here
shows
the
special
pattern
of
changes
in
the
net
surface
heat
flux.
So
we
found
that
generally
in
response
to
other
changes,
the
ocean
can
absorb
heat
from
the
atmosphere
generally
with
general
range
of
about
50
to
60
degrees
south
in
the
ocean,
in
pacific,
also
in
atlantic
and
there's
not
only
average
changes
in
net
surface
heat
thrust
is
shown
as
the
black
curve
in
panel
b
here.
N
So
we
found
that
the
southern
ocean
against
head
from
the
atmosphere
around
are
49
to
61
degrees
south,
but
lose
heat
over
certain
net
nine
and
the
earth
over
61
to
72
degrees,
south
and
many
heat
sizes
can
contribute
to
net
change.
So
here
we
find
that
the
old,
long
changes
induced
very
strong
short
wave
warming
to
a
source
about
55
degree
celsius
and
the
short
wave
warming
is
the
incompression
between
our
reduced
short
wave
reflection,
the
coral
curve
in
penalty
here
and
the
amplified
shorter
wave
reflection,
which
is
gold
curve
in
penalty
here.
N
The
reduced
short
wave
reflection
is
associated
with
an
active
cs,
retreating
as
well
as
estrogen
feedback.
Well,
the
gold
curve,
which
is
the
amplified
short
wave
reflection,
is
associated
with
changes
in
colors
and,
besides
the
short
wave
warming,
we
find
that
in
sensible
head,
which
is
a
green
curve
in
part
of
here.
The
sensible
head
generally
follows
the
patterns
in
net
surface
heat
flux
here,
and
the
sensible
head
is
altered
by
the
air
c
temperature
contrast
and
the
surface
wind
speed.
N
The
special
pattern
in
panel
c-
and
you
hear
another
average
impact
f
here
and
also
we
found
that
the
other
changes
lead
to
latent
heat
cooling.
The
purple
line
you
can
be
here
and
the
latent
heat
of
cooling
is
probably
due
to
the
changes,
the
wind
speed
to
the
source
of
48
degrees
celsius,
and
also
we
found
that
the
cs
can
directly
change
the
surface
head
flux,
which
is
orange
curve,
impeller
bishi.
N
It
leads
to
positive
anomalies
within
about
50
to
60
degrees
celsius,
negative
change
plus
also
of
60
degrees
celsius,
yeah,
with
the
results
in
the
surface
with
fast
and
ocean
circulations.
Here
we
want
to
link
change.
We
want
to
link
the
ocean
head
up
shake
under
the
old
and
the
ocean
circulation
induced
changes
in
better
than
ocean
heat
transport
convergence.
So
here
we
analyze
the
heat
budget
in
which
we
define
the
ocean
heat
storage
as
a
difference
between
ocean
heat
uptake
and
the
meredith
ocean.
Heat
transport.
Conversions
under
the
results
are
showing
panda
beach
here.
N
N
However,
due
to
the
effects
of
ocean
circulation
and
the
building
process,
which
may
transport
the
heat
to
lower
electric
tools,
which
is
the
red
curve
here,
so
that's
the
ocean-
head
storage,
chipset
low
latitudes
about
44
degrees
hours
here,
corresponding
to
the
changes
in
the
ocean
heat
strategy
event
that
has
also
induced
southern
ocean
warming,
which
is
dependency
here.
The
only
industrial
ocean
warming
peaks
around
45
degrees
celsius
and
comparing
the
odor-induced
warming
with
the
historical
warming
in
the
panel
issue
notice
that
the
counterpart
here
are
different.
N
With
the
changes
in
temperature,
we
further
analyze
the
ocean
head
content,
time
steelers.
We
want
to
know.
To
what
extent
can
the
odor
changes
contribute
to
the
ocean
head
contact
on
normally
in
the
southern
ocean,
and
the
panel
here
shows
the
specially
integrated
photos,
ocean
head
content
anomaly
since
1958,
and
the
red
cross
here
refers
to
all
the
induced
non-nominee.
N
While
the
blue
curve
here
refers
to
the
historical
anomaly
and
we
found
that
the
other
induced
ocean
helicopter
trend
accounts
for
about
22
percent
of
the
historic
ocean
content
and
anomaly
in
the
southern
ocean,
and
we
further
separate
the
flood
depths
into
the
operation
and
the
deep
ocean,
which
is
the
panda
beach
here,
we
found
that
the
only
indulged
ocean
high
content
anomaly
is
most
stored
in
the
upper
2000
meters.
There
yeah
to
better
understand
the.
Although
induced
temperature
change
here,
we
further
decompose
the
changes
into
specimens
a
hip
component.
N
The
specialist
component
refers
to
the
changes
on
density,
coordinate
which
refers
to
which
is
associated
with
changes
in
water
mass
properties,
while
the
heat
component
is
associated
with
vertical
movement
of
the
sp
nodes,
and
here
are
the
decomposition
and
the
panel
b
here
shows
the
specialist
component.
So
we
found
that
due
to
all
the
changes,
the
warming
trend
extends
down
equator
world
and
downward
a
from
surface
layer
at
65
degrees
celsius
and
the
warming
roughly
occurs
in
upper
500
meters.
N
I
to
the
north
of
60
degrees
celsius,
and
the
dependency
here
is
the
hip
component.
The
hem
component
here
is
associated
with
the
vertical
movement
of
the
aspect:
noise,
which
is
the
red
figure
here.
So
we
find
that
the
vertical
movement
of
the
aspignaus
is
most
dominated
by
the
surface
when
swiss
curve
anomaly.
So,
in
particular
the
positive
wind
stress
curve
anomaly
between
40
to
about
52
degrees
celsius.
N
It
depends
as
big
noise.
Therefore,
it
contributes
to
the
equatorial
world
and
vomiting
there
and,
however,
two
thousand
of
62
degree
cells,
the
negative
as
big
nose
and
the
negative.
Only
stress
curve
and
norm
is
shallows
as
peak
noise.
However,
due
to
the
vertical
temperature
invariant
in
this
region,
the
shallowing
speed
noise
will
lead
to
subsurface
warming
in
this
region,
and
here
are
some
of
the
conclusions.
N
The
shelling
has
beginners,
can
lead
to
subsurface
warming
and,
however,
until
north
of
50
degrees
celsius,
the
depletion
of
warming
that
christmas
to
the
ocean
histories
makes
is
primarily
due
to
the
deepening
big
noise,
and
we
also
found
that
the
large-scale
patterns
of
a
spindle
movement
are
consistent
with
the
surface
of
the
larynx
orange
juice.
Carbon
nominee,
which
identifies
the
important
role
of
wind
changes
in
the
southern
ocean.
Heat
redistribution
in
response
to
all
enforcing
yeah
and
all
the
results
are
prevalent
at
present
here
have
been
published
on
climate
dynamics.
A
Yeah,
just
at
the
last
point,
you
indicated
that
the
winston's
curl
is
playing
a
bigger
role
than
winstress
itself.
Why
is
that?
Oh
you
mean
this
one.
Well
in
your
last
conclusion,
you
have
the
fourth
one.
It
says
the
vince's
curl
anomaly
is
right,
yeah,
so
it's
not
the
zono
instance
itself,
but
it's
the
curl.
That's
what
you're
saying
right.
N
F
Actually,
all
right,
I
think,
we'll
move
on
to
our
last
talk
of
this
session,
we're
very
close
to
staying
on
time
here.
So
our
last
talk
will
be
from
kim
shearer.
The
title
is
marine
wild
capture,
fisheries
after
nuclear
war.
O
Great
thanks,
I'm
a
postdoctoral
researcher
at
the
autonomous
university
of
barcelona,
and
here
today
I
wanted
to
present
something
slightly
different.
I'm
going
to
shift
the
topic
to
catastrophic
and
sudden
climatic
change
and
to
fisheries
and
I've
squeezed
a
lot
of
things
in
here.
But
the
point
is
to
give
you
an
example
of
what
kind
of
questions
that
cesm
outputs
can
be
used
to
investigate
right,
but,
first,
why
is
it
important
to
think
about
the
consequences
of
nuclear
war?
O
This
figure
from
cpr
shows
the
number
of
nuclear
warheads
by
country
in
2019,
and
even
though
there's
been
a
large
decrease
in
the
number
of
weapons
since
the
cold
war,
there
is
definitely
enough
warheads
currently
to
cause
huge
destruction.
So
there's
an
also
an
increasing
number
of
countries
that
have
these
weapons,
for
example,
notably,
india
and
pakistan
have
rapidly
growing
weapons
arsenals,
and
these
two
neighboring
countries
have
been
in
armed
conflict
several
times
in
the
past
decades.
O
So
the
existence
of
these
weapons
means
that
a
large-scale
conflict
is
possible
and
beyond
the
direct
destruction,
a
nuclear
war
could
also
create
a
global
climate
disaster.
Detonations
of
weapons
in
urban
areas
could
cause
huge
fires
and
firestorms,
and
the
soot
from
these
fires
could
rise
into
the
upper
atmosphere
so
similar
as
during
large
volcanic
eruptions.
These
emissions
could
block
out
sunlight
and
cool
the
earth
for
several
years,
and
this
is
where
csm
or
other
earth
system
models
comes
in.
O
O
This
orange
chunk
here,
although
today,
marina
inland
aquaculture
combined,
is
equally
large
essentially,
but
this
orange
chunk
provides
about
10
of
the
total
animal
protein
production
globally
and
even
though
this
is
very
small
in
terms
of
calories
and
also
relative
to
other
animal
produ
production,
fisheries
could
be
a
pretty
important
food
source
if
land-based
production
systems
begin
to
fade.
O
So
I'm
going
to
start
a
little
bit
by
briefly
explaining
how
we
model
global
fisheries
and
cheryl
talked
about
this
this
morning.
But
there
are
for
those
of
you
who
didn't
hear
there
are
several
global
fisheries
models
that
use
gridded,
either
two
or
three-dimensional
environmental
variables
like
mpp,
sst,
oxygen,
etc,
to
model
the
abundance
of
fish
in
the
global
ocean
and-
and
this
figures
cheryl
also
showed
and
but
similar
just
like.
O
There's
the
c-mit
for
climate
models,
there's
the
fish
mip
for
marine
ecosystem
and
fisheries
models,
and
this
study
by
lutzen
colleagues
made
ensemble
model
projections
for
fisheries
and
the
climate
change
with
six
global
fisheries
models.
The
first
figure
here
shows
the
change
the
projected
change
in
total
global
fish
biomass
under
four
rcps,
and
these
lines
here
are
averaged
across
six
different
marine
ecosystem
models
to
the
right.
O
We
have
rcp
8.5,
where
each
model
projection
is
is
outlined
and
also
dashed
and
solid
lines
show
the
different
results
and
from
when
enforcing
with
gdp,
gfdl
or
ipsl.
O
This
is
a
quick
overview
of
boats
and
in
the
model
we
simulate
the
relationships
between
the
environment
to
the
left,
very
simplistically,
represented
by
net
primary
production
and
sea
surface
temperature,
the
growth
of
size,
structured
fish,
structured
fish
populations
here
in
the
middle
and
the
amount
of
fishing
that
humans
exert
here
to
the
right,
so
the
water
temperature
in
the
mpps.
O
So
two
of
the
nuclear
war
simulations
that
we
use
in
this
study,
we
used
six
different
nuclear
war
scenarios
developed
by
two
nadal
and
cupidal
and
they're
named
after
the
terragrams
of
soot
that
they
emit
essentially
representing
the
magnitude
of
the
war.
There
are
five
india,
pakistan
scenarios
and
one
for
war
between
the
u.s
and
russia,
and
here
to
the
right,
I
summarized
the
effect
that
these
sort
of
missions
would
have
on
the
radiative
forcing
and
the
sst,
and
these
are
maximum
annual
global
means
post-war.
O
So
my
colleagues
used
csm
to
simulate
the
impacts
of
these
war
scenarios
on
the
global
climate
and
we
simply
took
the
changes
in
sea,
surface
temperature
and
mpp
and
used
it
as
inputs
to
boats
to
estimate
the
impacts
on
fish
and
fisheries.
O
But
a
key
question
is
whether
people
would
fish
the
same
after
nuclear
war.
It's
a
pretty
serious
disruption,
socio-economic
disruption.
So
our
approach
here
was
to
again
set
up
a
few
very
simple
scenarios
for
the
potential
socioeconomic
responses.
First,
a
kind
of
control
scenario
where
fishing
goes
about
the
same
way
as
before
business
as
usual.
O
Second
is
possible
that
lack
of
food
on
land
would
intensify
the
search
for
fish,
so
we
have
an
intensified
fishing
scenario
and
it's
also
possible
that
the
war
would
limit
our
ability
to
fish,
maybe
due
to
lack
of
fool
or
fuel
or
destroyed
infrastructure.
So
we
have
a
decreased
fishing
ability
scenario
too.
O
All
right,
I'm
gonna
give
you
some
quick
results.
This
figure
here
shows
the
total
the
change
in
total
global
fish,
catch,
total
global
biomass
and
also
in
the
fishing
effort
relative
to
a
control
control
scenario,
without
war
for
the
different
for
the
six
different
nuclear
war
scenarios
from
five
teragrams
to
150
kilograms,
and
this
row
shows
the
business
as
usual
scenario.
So
here
we
essentially
essentially
only
see
the
direct
climatic
effect.
O
You
can
see
in
the
middle
that
the
decrease
in
biomass
is
up
to
about
minus
20
in
the
150
teragram
case,
which
in
turn
leads
to
about
a
30
decrease
in
the
catch.
But
you
can
also
see
that
in
the
five
teragram
case,
the
decrease
is
only
a
few
percent,
so
relatively
small,
especially
compared
to
crops,
and
this
row
here
shows
what
happens
if
we
have
intensified
fishing
due
to
the
war,
and
here
you
can
see
in
the
cache
that
we
can
get
a
temporary,
but
quite
small,
increasing
catch
up
to
about
20
percent.
O
If
the
climate
shock
is
not
too
large
and
then
intensified,
fishing
of
course
causes
a
rapid
decrease
in
the
remaining
fish
biomass,
which
is
the
reason
why
the
increase
can
only
be
sustained
for
a
few
years.
O
O
So
in
the
paper
we
also
investigated
the
difference
in
the
amount
of
emergency
fish
catches
that
we
could
potentially
get
if
all
fisheries
were
poorly
managed
to
the
left
or
versus,
if
they're
well
managed
before
the
war
here
to
the
right,
and
we
estimated
that
if
all
fisheries
were
well
managed,
catches
from
intensified
fishing
could
become
large
enough
in
the
first
year
after
the
war
to
replace
almost
40
percent
of
all
other
animal
production.
So
this
could
buy
some
time
to
adapt
food
production
systems
on
land.
O
So,
in
conclusion,
the
climatic
shock
causes
up
to
a
30
and
20
decrease
in
catch
and
biomass
respectively,
and
this
these
numbers
are
comparable
with
the
projected
end
of
century
decline
under
rcp
8.5,
and
it
is
the
light
reduction.
It's
a
key
driver
of
change.
Light
reduction
causes
the
decrease
generally
in
the
mpp,
which
translates
into
a
decrease
in
the
catch,
but
the
outcome
depends
a
lot
on
the
fishery
responses
and
also
on
the
pre-war
fisheries
status.
O
We
find
smaller
impacts
on
fish
than
on
crops,
but
we
have
to
remember
that
the
contribution
of
fisheries
to
global
calorie
intake
is
very
small.
However,
when
it
comes
to
animal
protein,
fisheries
can
probably
make
a
contribution,
and
we
find
that
well-regulated
fisheries
could
briefly
replace
about
forty
percent
of
all
other
protein.
O
So
that's
it
for
me.
There's
time
for
questions.
B
O
Right
so
I'm
gonna
most
likely
start
a
post-doc
with
sharon
harrison
in
by
christmas,
more
or
less,
and
then
we
are
definitely
gonna.
Look
at
the
effects
of
geoengineering
and
I'm
hopefully
going
to
start
a
little
bit
doing
it
and
but
yeah.
It
would
be
very
interesting
to
see
what
the
differences
are
and
the
similarities
yeah
thanks.
D
D
Okay,
my
question
is
the
nuclear
radiation
may
change
the
face,
the
gene
of
the
face,
so
the
face
may
grow
differently.
They
may
grow
like
a
bigger
or
smaller
and
better.
That
will
affect
your
result.
O
So
we
haven't
looked
specifically
at
the
direct
effects
of
radiation
fallout
on
fish.
O
That's
the
that's
the
question
right
yeah,
but
we
have
looked
at
other
studies,
so
there
are
examples
of
both
studies
that
have
been
made
when
during
nuclear
weapons
testings,
but
also
studies
looking
at
the
effects
of,
for
example,
the
fukushima
disaster
on
neighboring
fish,
and
to
my
understanding
I
don't,
I
haven't,
seen
anyone
kind
of
showing
evidence
for
changes
in
the
genetics
of
the
fish.
The
the
discussion
there
is
more
about
the
potential
for
humans
to
eat
the
fish
if
it
would
be
harmful
for
us.
F
All
right,
we'll
take
one
last
question
from
thomas
and
then
we'll
have
about
a
five
minute
break.
A
Hi,
thank
you
kim
very
much
for
the
interesting
talk.
I
was
wondering
that
the
suit
emissions
from
the
direct
consequences
of
of
bombardment
would
probably
occur,
and
your
maps
seem
to
confirm
that
in
the
northern
hemisphere,
but
when
you
show
and
and
you
say
that
the
main
driver
of.
O
So
I'm
I'm
not
an
expert
here,
but
from
what
I
understand
and
we
can
at
least
look
at
the
maps
from
what
I
understand
so
so
so
this
depends
on
the
scenario
for
the
u.s
russia.
O
Of
course,
we're
gonna
have
most
of
the
emissions
in
the
northern
hemisphere
for
india,
pakistan,
it's
gonna,
be
more
centralized
around
the
equator
or
very
localized
in
that
area,
and
I
from
what
I
understand,
the
simulations
suggest
that
soot
is
going
to
rise
high
enough
for
the
for
the
so
to
spread
into
the
atmosphere
and
cover
almost
the
entire
globe.
O
So
in
the
at
least
a
couple
of
years
after
there's
going
to
be
effects
not
only
close
to
the
actual
conflict
areas
but
around
the
earth,
so
I
think
that's
why
we
don't
see
any
clear
regional
effects
from
that.
I
don't
know
if
cheryl
can
is
she
there?
Maybe
she
has
something
to
add
about
this
yeah.
I'm
here
allen.
A
Is
here
allen
is
here
yeah
yeah,
that's
right!
The.
F
B
No,
I
think,
we'll
just
start
again
at
three
just
to
keep
on
time.
So
sorry,
it's
a
short
break.
O
I'll
have
to
switch
it
back
unless
you
want
to
be
the
one
that.
P
B
B
B
B
All
right,
hello,
it's
three
o'clock,
sorry
that
was
a
short
break,
we'll
get
started
with
the
the
final
session
of
our
working
group
meeting
before
we
have
a
discussion,
and
so
this
session
is
about
extremes
and
severe
weather
and
hydro
climate.
So
we'll
start
with
alyssa
stansfield
and
just
a
reminder
for
the
speakers.
I'll
give
you
a
two
minute:
warming
warning
at
eight
minutes
and
then
we'll
stop
it.
Ten
minutes.
J
Okay,
great,
so
thank
you.
My
name
is
alison
stansfield,
I'm
a
phd
candidate
at
stony,
brook
university,
and
this
is
work.
I've
been
doing
alongside
my
advisor
kevin
reed,
and
so
we
are
trying
to
use
simplified
cam
simulations
to
try
to
understand
the
response
of
tropical
cyclone
rainfall
to
climate
change,
and
so
it
probably
comes
as
a
surprise
to
no
one
here
that
predicting
changes
in
tropical
cyclone
rainfall
is
difficult.
J
There's
a
lot
of
things
that
go
into
predicting
the
amount
of
precipitation
from
tcs
at
a
given
location
such
as
the
number
of
landfalls,
the
translation,
speeds
of
the
storms,
their
sizes
intensities
and
the
precipitation
rates,
and
all
of
these
factors,
also
interplay
amongst
each
other
to
you
know,
produce
the
amount
of
rainfall
from
tropical
cyclones,
and
this
is
important
because
rainfall
from
tropical
cyclones
can
be
not
only
deadly
but
also
very
economically
destructive.
J
And
so
what
do
we
know
about
how
tc
rainfall
is
impacted
by
climate
change?
So
the
theory
kind
of
behind
this?
J
The
predictions
is
the
clausius
clapeyron
rate,
which
so
basically
the
thought
is
that
since
water,
the
amount
of
water
vapor,
the
air
can
hold
increases
by
about
seven
percent
per
degree
of
warming,
that
the
rainfall
from
extreme
events
like
tropical
cyclones
will
follow
that
rate,
and
so
the
consensus
from
different
modeling
studies
is
certainly
that
yes,
tc
rainfall
should
increase
in
the
future,
but
by
exactly
how
much
and
if
that's
going
to
follow
the
quasi's
clapper
on
rate
is
kind
of
uncertain.
J
So
we
kind
of
want
to
try
to
hammer
down
kind
of
this
uncertainty,
and
so
to
do
this,
we
want
to
use
simplified
cam
models.
So
again,
I'm
using
the
community
atmosphere
model
version,
five
with
the
official
rce
mip
comp
set
that
has
come
out
with
csm2
and
I'm
using
the
protocols
of
rce
mip,
which
is
which
are
detailed
in
wing
it
all
2018,
except
I'm
adding
rotation.
J
And
so
the
idea
is
that
you
know
we
start
with
the
real
earth.
We
can
represent
the
real
earth
using
like
amip
style,
more
realistic
simulations.
Then
we
can
simplify
further
to
the
rce
simulations
which
are
aqua
planets.
They
have
uniform
sst
everywhere.
The
solar
insulation
is
uniform
all
over
the
globe
and
the
aerosols
are
constant
and
so
again,
then
we
get
to
my
simulation,
which
is
like
rce
mip
except
the
earth
rotates,
and
we
are
running
the
cam
at
about
28
kilometer
resolution.
J
So
I
run
11
simulations
with
a
global,
uniform,
sst
varying
from
295
to
305
kelvin,
again,
I'm
using
28
kilometer
grid
spacing
and
the
sc
dynamical
core
and
so
quickly.
J
What
do
these
simulations
look
like
so
you'll
see
on
the
left,
an
animation
of
wind
speed
and
the
300
kelvin
simulation,
and
so
all
of
these
little
dots
here
are
tropical
cyclones
tracked
using
the
tempest
extreme
software,
and
so
we
see
what
we
see
is
that
tcs
in
these
tc
worlds
tend
to
form
in
the
tropics
and
subtropics
move
pole
word
with
beta
drift
and
then
kind
of
accumulate
up
near
the
poles,
and
so
we're
not
the
first
to
use
these
type
of
simulations.
J
But
what
I'll
just
note
here
is
that
for
the
precipitation
analysis
that
I'm
going
to
show
you,
we
limit
it
to
40
degrees
south
to
40
degrees
north,
because
we
kind
of
want
to
avoid
the
complications
of
the
tcs
that
are
near
the
poles
and
the
interactions
between
them.
We
don't
want
to
deal
with
that.
J
So
here's
just
some
tc
precipitation
composites,
so
the
the
each
box
is
different
simulation
with
a
different
sst,
that's
up
on
the
top
left
and
then
the
number
of
like
snapshots
that
go
into
these
composites.
It's
on
the
bottom,
and
so
what
we
see
here
is
that
definitely
in
the
inner
core,
the
tc
precipitation
increases
and
then
also
it
seems
that
as
the
sst
increases,
the
precipitation
field
seems
to
expand
a
little
bit
as
well.
J
And
so,
if
I
do
a
simple
estimate
of
the
percentage
precipitation
change
per
kelvin
change
in
sst,
we
get
about
7.37
per
kelvin
change,
which
is
probably
slightly
above
clausius
clapeyron.
Although
there's
a
kind
of
a
large
range
of
uncertainty
in
the
way
that
we
calculated
that.
J
But
so
next
we
want
to
ask
how
much
do
the
thermodynamic
versus
the
dynamic
factors
increase
affect
the
increasing
tc
precipitation?
So
we
kind
of
created
this
equation
that
estimates.
The
change
in
tc
precipitation
is
equal
to
change
in
the
moisture
availability,
which
is
the
thermodynamic
factor,
plus
the
change
in
tc
intensity,
plus
the
change
in
tc
size,
which
we
consider
the
changes
in
intensity
and
size,
the
dynamic
changes,
and
so
we
kind
of
divide
this
up
into
like
these
partial
derivatives.
J
J
The
dp
changes
in
based
just
on
the
changes
in
sst
change
in
intensity
and
change
in
size
are
estimated
using
a
poisson
regression
and
then
the
changes
in
intensity
and
size
per
change
in
ssd
are
estimated
with
linear
regressions
of
the
changes
in
mean
intensity
and
size,
and
so
these
are
the
kind
of
rudimentary
results
of
this
analysis.
So
we
look
at
two
different
precipitation
variables:
the
99th
percentile
and
the
accumulated
tc
precipitation,
and
so
the
colors
correspond
to
the
different
kind
of
contributions.
J
Well,
first
of
all,
the
the
total
change
in
precipitation
per
kelvin
is
8.58
for
the
99th,
percentile
precipitation
and
6.62
for
the
accumulated
so
they're
around
quasi-copper
on,
although
the
99th
percentile
is
certainly
a
little
bit
above,
and
what
we
found
was
that
the
sc,
so
the
thermodynamic
contribution,
kind
of
dominated
the
precipitation
increases
for
both
accumulated
and
99th
percentile.
J
J
So
just
some
quick
results
and
these
top
ones.
I
did
not
show
in
this
presentation,
but
I
just
thought
I'd
mention
them
in
these
simulations,
as
the
sst
warms
there's
less
tcs
at
any
one
time.
The
average
tc
is
more
intense
and
larger
and
there's
more
extreme
precipitation
rates
within
the
tcs
and
more
rainfall
is
coming
from
these
extreme
precipitation
rates
and
then,
when
we
dove
a
little
bit
further
into
the
tc
precipitation
analysis,
we
found
that
the
exact
results
kind
of
depend
on
the
precipitation
metric
that
you
use.
J
The
precipitation
changes
were
around
the
clausies
copper
on
rate
slightly
above
for
the
99th
percentile
and
the
thermodynamic
contributions
dominated
over
the
dynamic
contributions
to
the
change,
and
while
the
intensity
contributed
more
to
the
99th
percentile
change,
the
outer
size
changes
contributed
more
to
the
accumulated
precipitation
changes,
and
so
now
my
current
work,
which
I'll
just
briefly
mention
is.
I
want
to
link
the
idealized
modeling
back
to
the
real
world.
J
So
I
want
to
take
the
this
rotating
rce
mip
idealized
simulations
see
how
they
compare
to
amip
simulations
as
well
as
observations,
and
so
I
won't
go
too
much
in
detail
into
these
plots,
but
these
are
joint
dependence,
plots
of
99th
percentile
percept
tc
precipitation
their
dependence
on
sst,
which
is
the
y-axis
here
and
intent,
tc
intensity,
which
is
the
x-axis,
and
so
these
plots
well.
J
The
two
on
the
left
are
fairly
new,
so
I'm
still
kind
of
working
on
this
analysis,
but
so
far
it
seems
that
the
kind
of
pattern
of
these
precipitation
changes
are
consistent
between
the
idealized
simulation
and
the
amip
simulations,
and
these
observations,
which
I
got
from
the
emerge
satellite
but
I'll,
probably
be
using
other
observations
as
well.
J
So
I'll
just
mention
that
this
most
of
this
work
is
under
review
at
jgr
atmospheres
and
if
you
have
any
questions
or
want
more
information,
if
I
have
time
for
questions,
I
can
take
them
and
if
not,
my
email
is
up
on
this
slide.
Thank
you.
J
Yeah
that
could
definitely
be
kind
of
an
intermediary.
J
Another
kind
of
more
intermediary
thing
I'm
thinking
of
using
are
is
a
reanalysis,
so
I
think
I'm
going
to
use
the
era
5
reanalysis
to
do
because
that's
you
know
between
maybe
observations
and
amip
simulations,
I'm
not
sure
if
I'll
be
able
to
get
any
more
idealized
simulations,
as
you
mentioned
in
just
because
I'm
probably
finishing
my
phd
next
springs,
but
that
would
definitely
be
some
interesting
to
do.
B
Okay,
there's
a
question
in
the
chat:
what
what
is
it
by
the
simplified
model
that
makes
rainfall
fields
look
more
symmetrical
than
they
are
in
observations
and
fully
coupled
simulations.
J
That's
a
very
good
question.
My
first
guess
is
that
there
is
just
way
more
observations,
so
I
don't
know
if
you
are
way
more
like
snapshots
that
go
into
the
composites
that
might
be
making
them
more
symmetrical.
J
Additionally,
the
other
thing
is
that
there's
really
not
a
lot
of
wind
shear
in
these
simulations,
so
you
know,
there's
not
any
like
extra
tropical
transitioning
storms
to
kind
of
make
a
more
asymmetrical
shape
of
the
storms.
So
that's
my
my
two
thoughts.
That's
an
interesting
question,
though,.
Q
Q
I'd
first
like
to
take
a
second
to
thank
the
organizers
of
this
workshop
and
the
climate
variability
and
change
working
group
for
setting
this
up
and
giving
me
the
opportunity
to
participate.
Today,
I
will
be
sharing
with
you
some
of
our
work
on
the
impacts
of
climate
variability
on
large-scale
severe
weather
environments
over
the
united
states.
I
would
also
like
to
acknowledge
my
collaborator
collaborators
listed
at
the
bottom
of
the
slide
in
the
united
states
spring
marks
the
onset
of
the
severe
storm
season,
which
annually
accounts
for
billions
of
dollars
in
damage.
Q
For
instance,
this
figure
is
from
a
recent
paper
by
kristen
rasmussen
and
colleagues.
It
shows
the
end
of
century
average,
mixed
layer,
conductive
available
potential
energy
in
the
left,
column
and
conductive
inhibition
in
the
on
the
right.
For
the
months
may,
through
june,
they
force
dwarf
with
era
5
data
to
produce
a
control
run
shown
in
the
top
row
and
a
pseudo
global
warming
run
in
the
middle
row.
Using
multi-model
mean
changes
in
climate
from
rcp
8.5
cm5
simulations.
Q
The
difference
between
the
two
thus
can
be
viewed
as
the
force
signal
associated
with
climate
change.
Briefly,
what
you
notice
are
large
increases
in
both
the
end
of
century
convective,
available
potential
energy
and
convective
inhibition
inhibition
over
the
eastern
half
of
the
u.s,
with
the
largest
changes
over
the
southern
and
southeastern
regions.
Q
Q
Now,
instead
of
using
a
time-slice
approach,
as
was
previously
seen,
we
are
using
the
cesm2
large
ensemble
to
compute
severe
storm
parameters
directly
and
examine
how
they
have
changed
over
time.
For
instance,
shown
in
shown
is
the
cesm2
ensemble
mean
of
cape
from
1850
to
2100
in
the
dark
blue
line,
which
illustrates
the
forced
response
and
the
variability
of
individual
ensemble
members
are
shown
in
the
background,
in
light
blue
utilizing
a
large
ensemble
approach
from
the
cesm2.
Q
Q
Q
Previous
studies
have
shown
cape
so
6
to
be
useful
in
discriminating
between
severe
and
non-severe
storms.
Our
focus
is
on
the
march
to
june
seasonal
averages,
as
as
it
is
the
peak
severe
weather
season
for
the
united
states
to
validate
our
model.
We
are
using
various
read
analysis
products,
but
here
I
will
only
show
results
from
era
5..
Q
Q
Recalling
the
forced
signal
of
cape
adapted
from
rasmussen
and
colleagues,
we
can
see
similarities
in
the
patterns
of
end
of
century
cape
and
historical
cape,
so
6
trends,
especially
over
the
southeast
u.s,
although
when
you
consider
a
march
through
june,
keep
so
6
trends
from
era
5
for
the
same
period.
Seen
here
on
the
right,
there
is
a
significant
difference
in
the
patterns
and
magnitudes
now
transform
re-analysis
projects.
Q
Now
not
only
are
the
magnitudes
of
cape
so6
of
similar
values,
but
the
patterns
are
consistent
as
well
to
further
illustrate
the
run
to
run
differences
due
to
natural
climate
variability.
Here
I
show
the
cape
so
6
trends
of
all
14
individual
members
over
the
historical
30
year
period,
as
well
as
the
era
5
trend
for
the
lower
right
again
as
each
ensemble
member
is
a
single
realization
of
how
nature
might
have
evolved.
Q
The
era
5
falls
within
a
range
of
plausible
outcomes
over
this
period
simulated
by
the
cesm,
therefore
giving
us
confidence
in
using
this
model
as
a
tool
to
look
at
future.
Projected
changes
in
convective
environments,
as
I
will
show
next
shown
here,
is
the
evolution
of
keep
so6
over
the
next
30
years
from
2020
to
2049..
Because
of
time.
I
will
here
after
only
focus
on
cape
so6,
although
we
are
analyzing
other
parameters
as
well
as
a
reminder,
the
red
colors
here
indicate
increase
while
blue
indicates
a
decrease.
Q
This
figure
shows
all
eight
future
ensemble
members
from
the
cesm,
as
well
as
the
ensemble,
mean
all
forced
with
ssp
5
8.5,
forcing
in
the
bottom
right.
We
can
see
the
ensemble
mean
trend
or
the
forced
greenhouse
gas
response,
which
shows
increases
in
cape
so6,
almost
everywhere
over
the
continental
us,
but
particularly
in
the
southeast
and
over
the
gulf
of
mexico.
What
is
most
interesting
is
that,
although
the
ensemble
means
shows
increases
in
cape
so-6,
similar
to
historical
simulations,
there
appears
to
be
large
deviations
between
each
member,
for
example,
to
highlight
contrasting
runs.
Q
Another
aspect
of
my
graduate
work
is
to
focus
on
the
specific
regions,
to
quantify
the
role
of
internal
variability
and
to
further
understand
how
natural
variability
in
these
conductive
parameters
might
be
tied
to
regional
changes
in
sea
surface
temperatures,
although
I
am
evaluating
various
regions
such
as
changes
in
convective
environments
over
the
great
plains
or
the
south
central
us
today.
I
am
only
sharing
results
over
the
southeast
u.s
to
illustrate
the
relationship
between
cape
so-6
and
sst's
that
region
alone.
Q
Moreover,
as
I
pointed
out
in
my
previous
slides,
this
area
is
one
of
the
largest
variabilities
in
future
cape,
so
6
simulations
before
we
dive
into
any
of
the
sst
relationships.
This
time
series
depicts
the
march
through
june
ensemble
mean
of
cape,
so
6
trends
in
white,
as
well
as
the
individual
members
3
and
8,
which
show
different
trends
to
2049
over
this
region.
Q
Although
the
maximum
correlation
values
in
the
gulf
are
of
roughly
0.6,
the
spatial
pattern
is
still
very
coherent
to
further
evaluate
the
relationship
between
sst
and
southeast
uscape
s06.
We
computed
a
composite
analysis
of
those
for
those
years
in
which
the
absolute
magnitude
of
cape
s06
exceeded
one
standard
deviation.
Q
The
results
confirmed
that
the
years
when
the
model
simulates
an
increase
in
cape
so-6
correspond
to
years
of
above-average
sea
surface
temperatures
over
the
gulf
of
mexico.
On
the
other
hand,
when
the
model
simulated
decreases
in
cape
decreases
in
cape,
so
6
there
were
below
average
sea
surface
temperatures
in
the
gulf.
Q
To
conclude,
we
have
shown
that
cape,
so
6
is
projected
to
increase
over
the
united
states
in
both
magnitude
and
extent
through
the
century,
due
to
forced
climate
change,
with
the
greatest
increase
expected
over
the
southeast
u.s.
However,
over
the
next
several
decades,
changes
due
to
natural
climate
variability
are
equally
notable.
Q
J
Q
Now
I
know
that
I
am
approaching
my
time
limit
so
with
that
I
will
go
ahead
and
conclude
my
presentation
and
take
the
remaining
few
minutes
for
any
questions
or
comments
that
you
might
have.
Thank
you
again
for
your
time
and
attention
thanks.
B
A
Hello,
great
talk
thanks,
yeah,
really
interesting
topic.
I
was
curious.
I
know
you
mentioned.
You
didn't
have
a
lot
of
time,
and
so
you,
you
admitted
some
analyses,
but
when
you
shared
a
lot
of
results
for
kate,
the
product,
cape
and
zero
to
six
spoke
player,
sheer
I
think
it
was
yeah
and
yeah.
I
was
wondering
about.
Did
you
see
any
differing
trends
between
those
two
different
parameters,
because
there's
always
that
debate
about?
A
If
we
know
that
cape
is
increasing,
that's
a
pretty
robust
signal,
but
then
the
question
that
follows
is
that
that
there's
less
confidence
in
the
sheer
and
that's
so
relevant
for
severe
thunderstorms?
So
that's
my
first
question.
I
have
two
but
yeah,
so
I
was
wondering
if
you
saw
differing
trends
in
kate
versus
shear.
Q
Yeah
we
did,
as
I
showed
in
the
beginning
of
the
pre
presentation,
the
cape.
We
definitely
saw
a
positive
trend
over
the
southeast
u.s
gulf
of
mexico
region,
as
I
expressed
and
then
as
far
as
ss6
goes,
the
trend
was
a
little
bit
different.
It
was
definitely
a
little
bit
more
horizontal.
Q
I
guess
there
appeared
to
be
increases
all
the
way
across
the
continental
us
and
then
a
little
bit
more
of
a
decrease
below
over
the
gulf
of
mexico,
texas
region
and
across
those
latitudes,
but
we
are
looking
at
both
of
those
individually
and
overall.
When
you
look
at
the
product
of
cape
so-6,
it
does
appear
that
cape
is
the
main
parameter
driving
the
changes
between
the
product
of
so6.
A
That
makes
sense
and
then
just
a
question,
slash,
maybe
suggestion
I
was
wondering
about
the
motivation
for
the
analysis
over
water
for
severe
thunderstorms.
A
I
think
you
showed
some
of
those
correlations
and
also
the
environmental
results,
and
I
guess
my
suggestion
would
be
maybe
to
limit
it
to
land
just
because,
as
those
environments
are
affected,
over
land
they'll
be
modified
and
they're
blowing
out
your.
Q
Question
about
motivation
right
I
mean,
I
think
we
I
don't
think
we
were
really
trying
to
maybe
I
misspoke,
but
we
weren't
really
analyzing
any
of
the
severe
storms
over
the
water.
We
were
mostly
just
using
the
sea
surface
temperatures
to
correlate
them
with
the
storms
over
this
southeast
u.s,
so
any
of
the
the
water
that
was
included
in
that
analysis
in
that
southeast
u.s
region
was
masked.
So
I
didn't
actually
include
it
in
the
correlations
of
ssts
and
cape
ss6.
B
C
Good,
I
guess
it
just
matters
okay,
so
now
for
something
totally
different.
I'm
going
to
transport
everyone
into
the
cool
high
latitudes
for
those
of
us
who
are
suffering
in
heat
right
now,
so
I'm
going
to
talk
about
antarctica
and
atmospheric
rivers
and
a
little
bit
about
flavor
the
different
flavors
of
modes
of
variability.
C
This
presentation
is
in
collaboration
with
a
number
of
people,
michelle
mclennan
at
university
of
colorado,
jonathan
weil,
university
of
grenoble
and
arena
gordetskaya
at
university
of
aveiro,
okay.
So
just
just
a
quick
outline.
First,
I'm
going
to
try
to
motivate
you
and
tell
you
why
antarctic
atmospheric
rivers
are
so
interesting.
C
Show
you
a
little
bit
about
the
climatology
of
the
region
and
we'll
spend
a
little
bit
of
time
on
atmospheric
river
detection
tools,
because
this
actually
is
quite
important,
especially
for
regional
areas
such
as
antarctica
that
the
definition
of
a
of
how
we
define
an
atmospheric
river
matters
and
then
finally,
just
sort
of
launch
into
some
of
the
results.
I've
seen
with
different
modes
of
variability
and
what's
important
for
different
regions
of
antarctica,
all
right.
I
C
This
is
a
a
nice
little
movie
that
michelle
put
together.
It
shows
atmospheric
rivers
making
landfall
over
antarctica.
I've
highlighted
kuwait's
glacier
over
here,
because
I'm
going
to
sort
of
transport
you
over
here,
while
that
movie
is
going
to
this
figure
that
shows
snow
accumulation
totals
or
earth
weights
glacier
for
the
mean
period
of
the
era
5,
which
was
1979
to
2018
and
then
what
happened
in
2020.
C
Last
year
we
had
three
major
atmospheric
rivers
making
landfall
of
birth
weights,
one
in
february,
one
in
may
one
of
july,
and
these
were
massive
events
that
you
could
see
compared
to
them.
The
mean
really
did
increase
boost
the
snow
accumulation
totals
at
those
regions.
So
so
hopefully
this
will
help
to
convince
you
that
atmospheric
rivers
over
antarctica
actually
are
quite
important.
C
C
C
If
you
notice
the
the
color
bar
and
the
actual
range.
The
atmospheric
rivers
that
make
landfall
over
antarctic
are
actually
of
relatively
rare
events.
C
We
have,
on
average
three
days
of
atmospheric
river
activity
over
antarctica,
so
they're
relatively
rare
events,
but
when
they
do
occur,
they're
actually
quite
impactful,
and
the
other
thing
to
sort
of
take
away
from
this
plot
is
that
for
the
different
seasons
fall
winter
spring
and
summer,
you
see
different
hot
spots,
so
those
deeper
purple
colors
are
shipped
depending
on
what
the
season
is,
and
this
is
actually
quite
important
when
we
start
looking
at
modes
of
credibility,
all
right
so
on
to
a
little
bit
about
what
this
geographical
distribution
looks
like
here,
I'm
actually
plotting
a
whole
bunch
of
different
ar
dts.
C
These
are
atmospheric
river
detection
tools
from
the
atmospheric
river
tracking
method,
intercomparison
project.
So
there
we
have
quite
a
few
different
ones
to
look
at
all.
These
colors
mostly
are
ones
that
are
global.
A
C
General
detection,
algorithms
and
I
have
these
darker
purple
ones,
the
while
iwb
and
while
the
ibt
these
are
regional,
specific
and
specific
to
antarctica,
and
so,
just
by
looking
at
this,
as
you
can
see,
there's
actually
quite
a
dramatic
difference
between
what
a
regional
algorithm
will
tell
you.
What
a
global
algorithm
will
tell
you,
and
just
for
your
perspective,
here's
antarctica
over
here
where
this
is
zero
and
we're
actually
looking
at
180
around
the
horn
to
180.
C
the
only
place
where
there's
agreement
across
all
the
algorithms
is
actually
the
minimum,
which
is
this
frosty
area,
so
sort
of
making
this
a
little
bit
more
digestible
and
understandable.
I've
actually
grouped
all
the
global
ar
dts
compared
to
regional
ardt's
and
now
we're
looking
at
specific
regions
around
antarctica
in
terms
of
the
seasonal
cycle.
C
Here
is
again
my
little
cheat
sheet
map
of
antarctica.
Instead
of
using
the
antarctica
the
antarctic,
you
know,
land,
you
know
names
and
stuff.
I've
actually
used
the
the
ocean
that
it
backs
into
as
a
reference
just
for
just
just
because
for
me
this
is
understandable.
So,
apologies
to
all
those
antarctic
ice
sheet-
people
so
ea
would
just
be
the
east
antarctica
that
backs
into
the
atlantic.
C
A.I
would
be
atlantic,
india
and
so
forth
and
so
forth.
So
on
and
so
forth.
Wap
is
dis.
Essentially
I'm
looking
at
the
swiss
antarctic
peninsula
area
and
waa
is
this
backing
into
this
edmonson
sea
area.
So
I've
done
these
for
all
these
different
regions,
but
I'm
only
sort
of
highlighting
a
few
here-
the
west
antarctic
peninsula
and
the
amundsen
sea
area
versus
a
couple
of
the
east
antarctic
ones,
and
the
things
is
the
couple
things
to
pull
out
of
this
is
that
you
can
see,
especially
for
the
wp
and
the
ea.
C
You
know
it's
act,
absolutely
opposite
with
the
global
zoo
versus
what
the
regionals
do,
and
the
regional
ones
are
a
little
bit
more
closer
in
line
to
what
we
might
do
if
we
were
going
to
hand
detect
atmospheric
rivers
over
antarctica.
C
So
so
I
personally
view
these.
The
regional
ones
is
a
little
bit
more
accurate
than
the
global
ones,
and
I
can
sort
of
explain
that.
Well,
what
the
global
ones
are
doing
and
sort
of
picking
up.
You
know
basically,
the
southern
ocean
storm
track
as
opposed
to
atmospheric
rivers
that
are
actually
extending
into
the
continent
and
then
the
other
thing
to
sort
of
pull
away.
This
global
sent
to
actually
over
over
detect
ars
in
the
summer.
C
For
the
reasons
I
I
just
mentioned
all
right
so
now
I
sort
of
wanted
to
spend
the
rest
of
my
time
sort
of
talking
about
modes
of
variability
and
just
for
the
record,
I
guess
I
should
say
I'm
I'm
focused
on
these
regional
ar
dtt,
so
I'm
using
the
iwb
and
the
vibt
from
the
while
antarctic
specific,
because
I
feel
like
that
ar
dt
is
superior
for
this
type
of
analysis,
so
so
now
on
to
the
modes
of
variability.
Clearly,
the
the
big
one
is
the
sam.
I
don't
really
have
to.
C
I
think
explain
it
to
this
group
here.
So
this
is
just
sort
of
a
nice
little
graphic.
I
pulled
pulled
off
to
show
what
the
the
sam
is,
which
is
essentially
the
first
ufc
level
lt
level
pressure.
Here's
the
spatial
pattern,
I'm
actually
going
to
also
look
at
the
different
eofs
of
sea
level,
pressure
extending
into
ef1
2
and
3..
It
turns
out
that
uf
3
actually
has
some
significance
for
this
waa
area,
where
especially
earthquakes,
glacier
size
sort
of
also
want
to
include
that
and
what
I
show
you
next.
C
I'm
also
will
give
you
the
caveat,
because
the
time
limitation
started
trying
to
keep
track
of
my
time
here
that
I'm
going
to
only
just
show
you
some
results
on
west
antarctica.
I
do
have
them
for
east
antarctica,
but-
and
so
I'm
happy
to
talk
to
people
about
that
later,
but
for
now
just
sort
of
show
you
just
some
highlights
from
the
west
antarctic
piece.
C
So
this
is
looking
at
the
sam
and
the
psa2
phases
for
positive
phase
and
negative
phase
and
we're
looking
at
the
sea
level
pressure
which
is
the
color
contour
and
then
the
vector
arrows
are
the
is
the
moisture
the
low
level
moisture
flux.
So
this
is
this
is
uv
and
incorporated
into
reciprocal
water,
and
you
can
see
for
both
for
the
positive
phases
for
both
sam
and
psa2.
C
You
see
this
flux
of
moisture
into
this
west
antarctic
piece,
so
this
is
favorable
for
atmospheric
river
activity
and
for
the
negative
phases
you
see
this
outflow
from
the
continent
out
into
the
oceans,
which
is
not
supportive
of
atmospheric
river
activity.
So
this
is
sort
of
just
like
to
give
you
some
background.
Two.
B
C
Okay,
thanks
and
so
so
I
guess
we'll
just
jump
to
sort
of
the
the
these
different
modes
of
variability
regressed
onto.
I
did
both
a
number
of
variables,
we'll
just
sort
of
highlight
here,
the
the
precipitation,
because-
and
this
is
precipitation
attributable
they
are
so
now-
I'm
only
looking
at
ar
days
where
precipitates.
C
So
you
can
see
there
is
for
the
same
positive
phase.
You
have
this,
you
know
lower
pressures
and
it's
positively
correlated
with
precipitation
in
the
swiss
antarctica
piece
and
negatively
correlated
closer
to
the
ross,
the
rossi
area
for
psa2.
You
have
these
positive
precipitation
correlations
in
this
waa
section
of
west
antarctica,
which
is
where
earthquakes
glacier
resides.
So
this
these
are
and
they're
in
the
dotted
stippling.
That
implies
significance
for
these
progressions
all
right.
C
So
the
last
little
bit
here,
I
sort
of
started
looking
at
the
pdo,
which
is
the
first
thing
f
of
sst
anomalies,
but
instead
of
the
typical
thing
of
looking
at
over
the
western
u.s,
I'm
actually
going
to
draw
your
attention
here
to
west
antarctica,
because
there
is
some
correlations
with
west
antarctica.
C
So,
for
this
is
the
spatial
pattern
for
p.
For
the
pdo,
positive,
sorry,
the
I
you
know,
this
was
sort
of
a
quick
plot,
so
we
have
negative
ssd
anomalies
here
for
the
pdo
positive,
which
correlates
to
negative
correlations
with
precipitation
over
the
west
antarctic
piece
here.
So
if
you
flip
that
so
pda
pdo,
negative
phase
would
actually
be
supportive
of
ar
activity
in
the
west
antarctica
peninsula
and
over
the
west
antarctic
ice
sheet,
and
there
is
literature
out
there
that
has
just
come
out
that
actually
supports
this
okay.
C
So
just
quick
in
summary.
Hopefully
I've
convinced
you
that
atmospheric
antarctic
atmospheric
rivers
are
interesting.
I
certainly
think
they
are.
I've
showed
you
a
little
bit
about
the
climatology
and
hopefully
convinced
you
that
the
regional
ardt
that
I'm
using
is
the
most
appropriate
and
then
we've
looked
at
different
modes
of
variability,
thus
far
sam
psa2
and
the
pdo
and
their
influences
over
antarctica,
as
I've
done.
R
C
Of
subsetting
and
I'm
continuing
on
this
work
done
stuff
with
east
antarctica.
I
plan
to
move
forward
in
this
looking
at
model
resolution
on
the
variable
resolution.
Csm
that
adam
harrington
talked
about
a
little
bit
earlier.
I
think
a
number
of
other
people,
maybe
in
the
landice
working
group,
that
high
resolution
piece
over
antarctica
and
other
high
resolution
simulations
so
I'll,
stop
there
and
take
questions.
B
C
Yeah,
so
this
is
entirely
sort
of
dependent
on
what
region
you're.
Looking
at
it's
not
like
a
one-slice
fits
all
sort
of
a
thing
so
alice.
We
can
sit
down
and
look
at
different
maps,
and
I
can
show
you
you
know
which
pieces
are
better.
There's
there's
some
you
know
for
for
the
most
part,
like
you
know,
for
if
it's
well,
I
could
I
shouldn't
say
for
the
most
part,
but
there
is
evidence
that
there's
you
know
both
atmospheric
rivers
can
contribute
to.
The
these.
C
You
know
increase
the
snow
totals
like
we
saw
in
kuwait's
glacier
in
that
beginning,
but
also
an
east
antarctica
piece.
There's
evidence
that,
for
you
know,
sam
negative,
you
have
where
it's
warmer
and
not
necessarily
conducive
to
air
activity.
You
get
these
warm
air
intrusion,
events
would
actually
cause,
and
then
you
you
get.
You
know
something
that
may
result
look
like
an
atmospheric
river.
That
would
cause
a
melting
event,
because
it's
warm
air,
that's
doing
this
rain
on
snows
or
yeah
rain,
snow
sort
of
thing.
C
So
yeah,
I'm
not
sure.
If
that
answers
your
question,
I
hope
so.
B
A
P
All
right
so
yeah,
thanks
for
organizing
this
nice
session
isla
era
and
issue,
I
hope
you
can
see
the
screen
yeah.
So
this
is
a
it's
a
bit
work
in
progress
with
like
a
sort
of
a
illustrious
team
of
people
that
started
with
a
twitter
discussion
last
summer,
when
there
was
this
heat,
wave
or
sort
of
persistent
temperature
anomaly
over
siberia.
That
was
in
the
news
a
little
bit
and
I'll
talk
about
a
few
more
nuanced
aspects.
P
In
this
presentation.
The
relevance
in
general
was.
It
was
quite
interesting
in
the
context
that
it
was
just
a
very
strong
temperature
anomaly,
but
it's
also
coincided
not
just
last
year,
but
last
year,
in
particular
with
widespread
fires
over
siberia,
as
well
as
sort
of
worry
about
accelerated
permafrost
melts
during
these
really
persistent
temperature
anomalies,
and
so
we're
going
to
try
to
look
at
this
a
little
bit
in
more
detail.
P
P
I
was
one
involved
with
one
with
this
world
weather
attribution
team,
where
we
looked
on
the
one
hand,
that
sort
of
the
half
year
so
january,
through
june
temperature
anomaly
over
siberia.
You
see
the
map
of
this
plotted
here,
but
then
the
study
also
looked
at
the
daily
maximum
temperature
at
a
particular
station
which
went
over
100
degrees
fahrenheit,
which
was
the
first
time
in
the
arctic
circle,
so
that
really
grabbed
the
headlines,
but
was
maybe
less
interesting,
at
least
from
a
scientific
point
of
view.
P
Yeah
here
is
a
the
daily
temperature
time
series.
It
was
just
very
extreme,
but
it
was
an
extreme
event
in
general.
Also,
if
you
look
at
sort
of
the
half
year
average-
and
the
attribution
study
may
be
unsurprisingly,
then
concluded
that
at
least
this
first
half
year
was
effectively
almost
impossible
without
human
influence.
P
This
derives
from
sort
of
this
yeah
sort
of
expansive
analysis
of
observations
and
different
models,
large
ensembles,
to
come
up
with
a
probability
that
this
event
could
occur
in
today's
climate,
comparing
it
to
a
pre-industrial
climate,
and
so
the
probability
that
this
could
occur
was
without
climate
change
was
extremely
low.
That
led
to
this
headline
of
almost
impossible
people
also
looked
at
sort
of
the
the
drivers
again
in
particular
of
the
winter
and
the
spring
part
of
this
event.
P
A
study
by
overland
wang
pointed
out
that
yeah,
the
arctic
oscillation
was
very
positive
during
the
winter
2019
2020,
which
was
like
driven
by
a
strong,
very
strong,
polar
vortex,
like
I
think,
even
record
breaking
ozone
levels
in
the
arctic
and
then
sort
of
a
different
pattern
towards
spring
and
summer.
So
these
are
your
potential
height
anomalies,
so
you're,
seeing
strong,
negative
anomalies
aloft
in
the
early
part
and
then
more
positive
anomalies
and
more
closer
to
the
surface.
P
It's
also
exceptionally
well
forecasted
event,
I'm
not
going
into
that,
but
these
are
six
different
sort
of
seasonal
forecasting
models
and
in
in
several
of
them.
This
event,
which
is
this
the
the
skill
of
the
forecast
is
given
with
this
dashed
line
compared
to
like
sort
of
a
hind
cast,
was
very
well
among
the
best
forecasted
winter
sort
of
yeah.
C500
anomaly
correlations,
so
it
was
like
a
strong
event
and
well
forecast.
P
So
then,
then,
there's
a
little
bit
of
like
sort
of
my
personal
story.
I
started
looking
into
that
together
with
a
bunch
of
these
people
from
the
attribution
group,
and
then
I
started
a
new
job
at
cornell
university
in
fall
and
really
didn't
get
too
much
in
the
beginning,
and
so
the
whole
year
2020
passed,
and
it
turned
out
that
also
the
second
half
was
relatively
extreme.
P
So
I
decided
to
pivot
a
little
bit
and
focus
on
the
whole
year
and
look
at
what
caused
this
very
persistent
temperature
anomaly
over
the
whole
year.
So
here
are
daily
temperatures
over.
This
is
the
same
box,
I'm
just
looking
at
land
in
this
box,
the
same
box
as
was
used
in
the
earlier
attribution
study
and
again
you
can
see
that
this
really
compared
to
climatology
that
is
1951
to
1980.
P
P
This
is
just
plotted
again
here
against
yeah,
a
multi-modal
mean,
doesn't
really
matter
which
ones
seem
at
five
to
be
this
collection
of
hard
ensembles
or
whatever.
It's
generally
consistent
with
that
long-term
trend.
What
I
want
to
point
out
here
is
again:
these
are
the
monthly
mean
anomalies
over
the
region
of
interest
for
2020.
P
So
really
this
like
consistent,
positive
anomaly
throughout
the
year,
at
times
three
standard
deviations
over
climatology
and
then
in
light
gray,
you
see
all
the
other
years
in
in
the
observation
record,
so
we
wanted
to
these
apart
a
little
bit
more
to
understand
how
much
of
this
anomaly
was
due
to
internal,
mainly
atmospheric
circulation
variability
and
how
much
might
be
driven
by
other
by
other
factors
and
potentially
feedbacks,
and
we
applied
two
dynamical
adjustment
methods.
P
I'm
not
going
to
go
into
the
details
very
much,
in
particular
the
first
one
I've
used
extensively
and
presented
on
earlier.
That
was
coming
out
of
work
from
clara
dessert
and
lo
ontario
and
then
another
one
by
sebastian
simple.
P
The
point
really
here
is
to
estimate
with
various
methods,
the
contribution
of
atmospheric
circulation
and
its
variability
to,
for
example,
temperature
anomalies
in
this
case
and
try
to
separate
that
that
out,
and
so,
if
we
do
that
with
both
of
these
methods,
we
come
up
with
an
estimate
of
the
contribution
from
atmospheric
circulation
variability
to
well
the
entire
time
series,
but
more
specifically,
this
particular
year
and
sort
of
consistent
with
our
expectation
that
that
a
strong
architect
oscillation
is
strong,
polar
vortex
influenced
this
event.
P
We
see
a
large
contribution
from
atmospheric
circulation
variability
in
the
first
half
of
the
year
and
then
less
later
on
and
the
receipt
will
effectively
worse.
We
can
then
use
to
interpret
more
as
as
a
thermodynamic
driven
component,
and
we
see
that
this
was
the
one
that
dominated
the
temperature
normally
in
the
second
half
so
somewhat
contrasting
evolution
throughout
the
year.
P
I'm
not
going
to
go
into
too
much
detail,
but
you
can
go
in
and
look
at
the
actual
circulation,
anomaly
and
and
temperature
anomaly
in
each
month,
january
february,
etc,
and
then
the
dynamical
contribution
and
the
residual
throughout
the
year.
I'm
going
to
delve
into
that
here,
but
that's
something
you
can
get
out
of
these
methods
and
and
then
it
turned
out
that
like
right
around
the
time,
so
I
think
ultimately
published
in
in
yeah
in
spring
2020,
a
paper
by
wu
and
chen.
Looking
at
the
factors
that
lead
to
persistent
air
temperature
anomalies.
P
They
were
just
focusing
on
going
from
winter
to
spring,
but
over
mid
to
high
latitude
eurasia
and
they
sort
of
compiled
a
nice
recipe.
If
you
will
of
some
of
the
ingredients,
they
think
are
important
and
they
were
hypothesizing
based
on
yeah
earlier
work
by
others
that
particular
north
atlantic
ssts
are
favorable
for
a
persistent
arctic
oscillation.
In
this
case,
it's
not
always
clear
that
the
strong
polar
vortex
is
driven
by
that,
but
we
definitely
have
both
of
these
patterns
in
place
and
then
also
the
scandinavian
pattern
as
a
circulation.
P
Anomaly
pattern
becoming
important
in
spring
and
then
also
they
touched
upon
the
potential
for
sea
ice
anomalies.
But
this
was
sort
of
like
in
in
the
general
sense,
but
they
had
done
this
analysis
on
the
historical
record
without
looking
at
2020
yet,
and
so,
if
you
do
look
at
these
factors,
the
north
atlantic
sst,
it's
a
triple
pattern.
I'm
not
going
to
show
it
here,
but
particular
index
was
slightly
positive
throughout
winter.
Here
we
see
this
very
strong
arctic
oscillation
that
I
mentioned.
That
then
goes
back
to
normal
values.
P
Scandinavian
pattern
also
quite
high
in
in
certain
instances.
So
all
of
these
ingredients
were
in
place,
and
we
also
know
that
these
ingredients
are
important
for
for
these
linkages,
and
so
here
I'm
showing
the
correlation
for
each
month
between
atmospheric,
the
arctic,
oscillation
and
temperature
over
the
region
of
interest,
and
now
what
we
can
do,
that's
running
black
now
what
we
can
do.
P
We
can
look
at
this
also
in
the
dynamical
part,
the
blue
part
of
the
temperature-
and
we
do
see
that
this
variability
in
generally
is
very
strongly
correlated
with
with
the
arctic
oscillation,
while
the
sort
of
the
residual
is
more
thermodynamic,
part
isn't
correlated
just
as
we'd
expect
slightly
weaker
correlations
for
the
scandinavian
pattern.
P
Yep
and
then
we
looked
at
a
few
other
patterns.
P
Sorry
other
potential
drivers,
especially
in
context
of
this
second
half,
where
we
think
that
the
thermodynamic
part
is
more
more
important
like
the
siberian
snow
cover
which
was
exceptionally
low
during
2020,
then
soil
moisture,
which
also
had
a
almost
a
record
minima
in
in
june
and
then
sea
ice
along
the
siberian
coast,
which
also
was
record
low
in
in
2020,
and
so
those
those
factors
are
generally
also
correlated
negatively
with
temperature
in
in
parts
of
the
month
parts
of
the
year.
P
And
again
we
see
that
this
time
around.
If
you
do
this
dynamical
adjustment,
we
see
that
the
residual
temperature,
so
something
that
we
might
expect
to
be
more
thermodynamically
driven,
is
more
strongly
correlated
with
with
snow
cover,
soil,
moisture
and
sea
ice.
P
I
forgot
to
mention
in
all
these
cases,
temperature
is
deep,
so
we're
like
subtracting
the
long-term
temperature
effect
which
would
inflate
some
of
these
correlations
a
little
bit.
So
those
are
significant
in
any
case,
and
so
then
we
can
go
in
and
try
to
reconstruct,
especially
this
thermodynamic
part
of
2020,
with
a
regression
model
that
just
uses
snow,
soil,
moisture
and
sea
ice,
and
we
can
do
a
reasonably
good
job
to
reconstruct
this.
P
This
red
curve,
except
in
november
2020,
which
still
causes
some
headache,
because
a
very
strong
thermodynamic
contribution
suggested
by
dynamical
adjustment.
But
it's
not
easily
reconstructed.
Here's
the
generally
explained
variance
of
this
model.
It's
also
not
that
high
in
november.
So
some
some
questions
remain
in
this
part,
and
then
I
forgot
to
mention
this
like
I
have
a
whole
bunch
of
slides,
because
I'm
sort
of
working
on
a
larger
yeah
trying
to
get
up
to
speed
with
this
paper.
P
I
gotta,
unfortunately
skip
over
those
for
time
reasons
and
go
to
the
conclusions.
To
summarize
this
yeah
we
had
these
strong
temperature
anomalies
that
mainly
grabbed
the
headline
in
the
first
half
of
the
year,
but
continued
on
and
are
quite
interesting.
P
I
didn't
get
to
the
part
where
we
looked
at
the
model-
validation,
which
is
quite
challenging,
because
this
is
sort
of
an
unprecedented
event
and
like
to
study
the
actual
processes
which
isn't
usually
done
in
classic
attribution
studies.
P
It
becomes
quite
tricky
because
it
it's
in
any
in
a
lot
of
these
observational
metrics,
it's
it's
the
most
extreme
event,
but
generally
we
do
find
the
models
to
reproduce
these
mechanisms,
and
we
think
this
can
be
expanded
to
include
other
other
impacts
as
such
as
fire
or
thermo
permafrost
dying,
and
I
did
want
to
plug
this
announcement
for
a
workshop
that
does
look
at
attribution
specifically
of
multi-animal
to
decadal
changes,
which
is
a
bit
of
a
new
new
topic
and
we're
in
particular,
looking
for
people
who
work
with
with
large
ensembles
and
some
of
the
tools
that
you've
all
presented
on
it's
going
to
be
virtual
and
free
and
registration
deadline
is
coming
up
soon.
P
B
P
P
Pattern
correlation-
I
think
that
is
good,
so
I
I'd
have
to
look
into
that
or
maybe
easier
to
ask
these
authors
yeah.
I
think
that
was
simon
lee
and
the
other
paper
was
by
zach
lawrence
yeah.
B
B
R
And
so
hi
everyone,
I'm
a
postdoctoral
research,
fellow
at
washington,
state
university
vancouver.
So
I
will
present
some
of
my
work
I
have
been
doing
in
in
postdoc,
so
my
presentation
to
broadly
focus
on
understanding,
simultaneous
droughts
and
their
drivers
and
historical
and
future
climates.
So
I'd
like
to
acknowledge
my
collaborator
listed
here
for
their
suggestions
and
feedbacks
on
this
yesterday.
A
R
So
the
compound
droughts
events
are
defined
when
multiple
regions
experience
drought
simultaneously.
So
the
motivation
of
this
work
comes
from
a
historical
drought
event
which
occurred
during
1876
to
78
when
multiple
regions,
which
are
highlighted
with
yellow
and
pink
color
experience
the
drought
simultaneously.
R
So
it
was
a
multi-year
drought
which
caused
widespread
crop
failure
and
further
catalyzed
the
global
famine,
which
resulted
in
more
than
50
million
fatalities.
So
this
event
was
considered
one
of
the
worst
feminine
human
history.
However,
the
impact
from
such
event
may
not
be
same
in
today's
world,
but
still
it
may
have
a
serious
impact
on
global
food
system.
For
example,
this
kind
of
events
may
affect
the
food
production
and
which
may
increase
the
global
food
food
prices.
R
So
from
this
this
point
onward,
I
will
use
the
compound
route
instead
of
saying,
like
spatially
compound
routes.
R
R
R
So
we
study
the
compound
routes
over
these
stand.
Regions
mark
with
red,
solid
lines.
So,
basically,
several
of
several
of
these
regions
are
influenced
by
global
monsoon
systems,
and
these
regions
includes
important,
broad
bread,
baskets
and
vulnerable
populations
and
that
depend
on
ran,
fed
agriculture
for
their
their
livelihood.
So
within
each
region
we
consider
the
only
area
which
shows
the
high
variability
during
summer
monsoon
precipitation,
which
is
indicated
with
teal
color
text
and
red
color,
shows
the
fractional
area
with
higher
variability
within
each
astrox
region.
R
So
we
also
consider
sea
surface
temperature
over
which,
over
various
ocean
areas
marked
with
dash
blue
line
to
define
the
modes
of
natural
variability.
So
we
examine
the
influence
of
these
four
variability
modes
on
compound
route
characteristics.
So
these
modes
are
elino
tropical,
north
atlantic
atlantic
nino
and
indian
ocean
dipole.
R
R
So
we
define
our
region
under
drought
if
drought
affected
area
exceeds
a
80th
percentile
of
long-term
average
drought
area
and
we
define
the
compound
route
event
if
at
least
three
regions
out
of
10
experience
drought
simultaneously,
so
we
identify
the
specific
phase
of
four
variability
modes
relevant
to
compound
route
based
on
the
observational
analysis
and
due
to
larger
sample
size.
So
we
use
the
c
assembly
industrial
simulation
here
to
isolate
the
influence
of
each
individual
modes
and
their
occurrences
phase
on
compound
routes.
R
So
this
figure
shows
the
distribution
of
number
of
number
of
regions
experienced
route
simultaneously
during
the
occurrences
of
individual
and
combined
variability
mode,
which
is
indicated
on
the
on
the
x
axis.
So
the
number
on
top
of
each
box
shows
the
probability
of
compound
rounds
during
each
condition.
So
we
define
the
neutral
here
when
the
ssd
is
over
tn-
and
you
know
3.4
atlantic
nino
and
indian
oceans
are
are
within
a
plus
minus
0.5
hd.
R
So
the
horizontal
dash
line
shows
the
threshold
used
to
define
the
compound
routes,
so
the
significant
difference
in
the
distribution
is
marked
with
the
with
arrows,
so
the
distribution
of
drought
region
during
negative
phase
of
atlantic
nino,
so
which
is
like
a
negative
atlantic
and
the
negative
dna
and
during
alien
conditions,
are
significantly
significantly
higher
than
that
during
the
neutral
condition.
R
So
more
specifically,
the
probability
of
compound
route
increase
more
than
two
times
during
negative
atlantic
nino
and
tna
conditions
and
around
three
times
during
alien
new
conditions
related
to
the
neutral
conditions.
R
So
further,
the
co-occurrence
of
negative
tna
conditions
with
el
nino
increase
the
probability
of
compound
routes
nearly
seven
times
relative
to
neutral
conditions
and,
more
than
two
times
relative
to
alienal
conditions
alone.
So
we
find
the
similar
influence
of
individual
and
combined
modes
of
variability
on
distribution
of
drought,
area
and
drought
intensity.
So,
overall
we
find
that
the
the
overall
we
find
that
alino
is
the
major
driver
of
compound
routes
and
dna.
Negative
conditions
has
a
largest
amplifying
influence
on
the
new
dr1
compound
routes.
R
So
next
we
have
now
we
have
some
understanding
about
the
compound
route
and
its
driver.
So
we
further
extend
this
works
to
understand
the
compound
routes
in
future
under
business
as
usual
scenario,
rcp
8.5.
R
So
for
that
we
use
the
cscsm
large
assemble
for
this
analysis,
so
this
figure
shows
the
distribution
of
drought
reason
in
the
historical
climate
which
is
defined
as
1971
to
2000
and
the
future
climates
with
the
red
current,
which
is
defined
over
2071
to
2100.
R
So
the
number
on
top
of
each
box
shows
the
probability
of
compound
routes
and
the
horizontal
line
is
same,
which
shows
the
threshold
used
to
define
the
compound
route.
So,
overall,
we
find
a
significant
increase
in
the
distribution
of
drought
reason
in
future
relative
to
baseline
period.
That
is
the
kind
of
historical
period.
R
So
so
the
probability
of
compound
drops
increases
by
around
60
percent
in
the
future
relative
to
historical
period.
We
also
find
the
similar
changes
in
extent
and
severity
of
compound
drops
in
future
relative
to
climate,
historical
climate.
So,
specifically,
nearly
three
out
of
four
compound
droughts
in
the
future
are
classified
as
a
severe
compound
drought.
R
So
since
enso
is
found
to
be
a
major
driver
of
compound
drought
in
the
in
the
previous
analysis,
so
we
examined
the
historical
and
future
characteristics
of
enzo
index
and
their
association
with
the
compound
routes,
so
the
in
the
panel
a
so
that
shows
the
pdf
of
the
enso
index,
which
shows
almost
22
percent
increase
in
the
frequency
of
el
nino
and
linear
conditions,
which
is
consistent
with
the
previous
studies,
such
as
sky
at
all,
2014
and
15..
R
So
the
panel
b
indicate
the
distribution
of
compound
drought
associated
with
a
linear
linear
conditions
and
and
non-enso
drivers.
So
the
number
on
each
bar
shows
the
count
of
compound
route
associated
with
the
I
linear
and
laline,
and
non-answer
driver.
So,
although
alino
is
the
associated
with
the
major
fraction
of
the
compound
route,
la
lina
are
also
associated
with
a
significant
fraction
of
compound
routes
and
the
historical
climate
and
it
further
increase
in
the
future.
R
However,
the
reasons
they
affect
may
differ
so
overall,
the
frequency
of
compound
drought
associated
with
a
linear
linear
conditions
increases
by
70
percent
in
future
in
response
to
just
22
percent
increase
in
the
frequency
of
henson
condition
and
future
warming.
So
in
future,
nearly
three
out
of
four
compound
droughts
are
associated
with
and
so
and
so
conditions.
R
So
we
further
estimate
the
population
and
agricultural
and
exposure
to
compound
droughts
to
panel
a
shows,
the
comparison
of
agricultural
and
exposure
to
moderate
and
severe
compound
drought
between
historical
and
future
climates,
so
the
x
axis
shows
the
agricultural
land
exposure
to
moderate
and
civil
compound
route
and
historical
and
by
access
source
for
the
future
climate.
R
So,
however,
I
will
just
focus
on
the
exposure
to
the
severe
compound
routes,
so
the
agricultural
and
exposure,
so
so
the
agricultural
and
exposure
to
sewer
compound
routes
increase
almost
10
fold
in
the
future
climate
related
to
historical
climate.
If
we
compare
with
the
one
ratio
one
line,
so
similarly,
we
see
minutes
yeah.
Thank
you
yeah,
so
we
consider
so.
Similarly,
we
see
a
remarkable
increase
in
the
population
exposure
to
severe
compound
hard
route.
R
So
we
consider
the
population
exposure
under
all
ssps
to
highlight
the
difference
in
the
projected
population
scenario.
However,
ssp
5
is
the
largely
consistent
with
the
rcp
8.5.
R
So
under
ssp
5
scenario,
we
find
that
on
an
average
10
million
people's
per
year
are
exposed
to
severe
compound
route,
where
this
number
increased
220
millions
like
like
around
120
million
per
year
in
the
future
climate.
R
So
I
will
quickly
summarize
the
results,
so
we
find
that
el
nino
is
the
major
driver
which
control
the
largest
fraction
of
historical
compound
drought.
However,
the
co-occurrence
of
gold
ssds
over
tna
has
largest
amplifying
influence
on
a
linear
one
compound
route.
R
So
the
probability
of
compound
routes
increases
by
60
percent
in
the
future
climate
relative
to
historical
period
and
enso
related
compound
droughts
increases
by
70
percent
in
response
to
just
22
percent
increase
in
frequency,
offensive
conditions
and
future
warming,
so
the
so
the
increase
in
future
compound
drought,
corrections
result
in
up
to
almost
10
fold
increase
in
agriculture
and
land
population
exposure.
R
So
these
results
have
several
implications
for
wide
range
of
climate.
Sensitive
sectors
such
as
global
food
security,
insurance
industry
and
also
these
results
have
a
broad
implication
in
the
global
virtual
water
trade
network,
which
involve
water
intensive
goods.
Thank
you
I'm
happy
to
take
answer
and
also
we
are
revising
the
second
manuscript.
So
I
would
appreciate
any
any
feedback.
B
R
Yeah
so
initially
we
consider
the
u.s
southwest
analysis,
but
we,
if
we
consider
only
region
which
shows
the
high
variability
in
the
boreal
summer
precipitation.
So
we
found
a
really
small
fraction
of
area
which
shows
the
high
variability
over
this
regions,
and
we
use
the
criteria
when
at
least
30
percent
of
that
region
should
shows
the
high
precipitation
variability.
So
that's
why
we
excluded.
E
Hi
thanks
for
the
interesting
talk,
I
was
wondering
if
you
can
comment
on
how
model
dependent
the
results
are.
If
you
look
at
different
large
ensembles,
you
get
quite
different
future
projections
of
enso,
so
I
was
wondering
how
that
would
affect
this
study.
R
Yeah,
so
I'm
revising
that
paper
and
we
got
the
same
comment
from
the
reviewer.
So
I'm
working
on
that.
So
I
smile
couple
of
more
models
from
the
smile,
and
so
we
found
like
a
quite
good
consistency
in
terms
of
compound
route
characteristics
across
the
model,
but
definitely
there
are
some
differences
in
the
answer
response
in
different
models.
So
we
are
still
working
on
that
yeah.
R
So
this
figure
is
showing
like
a
same
like
a
distribution
of
number
of
drought,
region,
route
area
and
route
intensity,
and
so
this
is
the
first
three
box
plots
for
csm
and
the
middle
one
is
for
the
kanye
sm
and
last
three
one.
So
we
see
a
like
a
strong
consistency
between
these
models
in
terms
of
drought,
characteristics.
B
Okay,
thank
you
so
we'll
move
on
to
the
last
speaker
of
the
session
zhao
shin
bam,
talking
about
the
asymmetry
of
annual
stream
flow
responses
to
warming
in
the
western
u.s.
B
M
Screen:
okay,
good
afternoon,
everyone,
I'm
joshimba
from
the
department
of
geography
at
ucla.
Today
I
will
present
our
recent
work
on
understanding
the
symmetry
of
annual
stream
flow
responses
to
seasonal
warming
in
the
western
u.s.
More
than
26
percent
of
the
global
land
area
depend
on
no
male
as
their
dominant
water
resource.
Western
u.s
is
a
typical
snow-dominated
region,
which
snowpack
has
experienced
a
significant
decrease
under
climate
warming.
It
significant
decrease
in
snowmelt-generated
runoff.
These
changes
are
challenging
to
the
region's
social
economical
safety.
M
Therefore,
it
is
important
to
understand
how
and
why
the
stream
flow
changes
under
climate
warming.
The
climate
warming
signal
in
the
western
u.s
is
not
evenly
distributed
on
a
sub-annual
scale.
The
warming
in
a
warm
season
from
april
to
september
is
substantially
larger
than
warming
in
the
crew
season
from
october
to
march.
M
M
This
adult
did
not
explicitly
explain
why
the
symmetry
differs
and
their
results
are
from
only
one
hydrological
model,
so
whether
this
result
is
consistent
across
models
is
still
a
question
motivated
by
these
gaps.
We
ask
two
questions.
First,
whether
or
not
the
symmetry
of
shrimp
responds
to
signal
warming
is
motor
dependent.
Second,
why
the
annual
stream
flow
responses
to
seasonal
women
defer
across
basins?
M
You
have
a
new
draft
and
a
review
iwrr
will
also
expand
analysis
to
the
entire
western
u.s
at
a
much
finer
granularity,
and
I
will
use
some
of
its
results
in
this
presentation
too.
To
answer
the
questions.
We
use
four
land
surface
models,
including
wearable
infiltration
capacity
model,
no
np,
cachement
and
ssc
sma
at
1,
16
degree
resolution
we
fix
precipitation
and
perturb
the
daily
maximum
and
minimum
temperature
and
the
forwarding
scenarios
to
isolate
the
effect
of
temperature
on
stream
flow.
The
first
scenario
is
baseline.
The
second
is
three
degree:
warming
in
a
warm
season.
M
M
here
are
the
multimodal
results
of
streamflow
response
from
the
first
to
the
third
rows.
Are
stream
flow
responses
to
annual
warm
season
and
cruising
warming
within
each
subplot
from
top
to
down?
Are
responses
at
annual
warm
season
and
cruising
time
scale?
The
gray
bars
are
multi-modal,
mean
and
color
bars
are
for
different
models.
M
We
can
see
that
despite
the
absolute
values
defer
the
models
share,
similar
relative
magnitude
and
sign
and
a
multimodal
mean,
also
shows
stronger
responses
to
warm
season,
warming
in
a
colorado
and
columbia
river
basins,
while
the
other
two
basins
has
very
weak
or
even
reversed
symmetry.
So
here
we
answer
the
first
question.
The
stream
flow
response
and
symmetry
are
consistent
across
different
land
surface
models.
M
Now
we
answer
the
second
question:
why
asymmetry
defers
the
cross
basins?
We
apply
a
long-term
water
balance
framework
where
the
annual
evapotranspiration
change
equals
the
annual
runoff
change.
Therefore,
we
can
explain
the
annual
runoff
is
changed
back
explaining
the
annual
eruptions
variation
change,
and
this
makes
it
easier
than
directly
analyzing
runoff.
Since
the
evapotranspiration
is
a
more
instant
and
local
process
than
runoff
in
the
following
slides,
you
will
see
some
subscripts.
The
first
subscript
indicate
the
season
and
the
second
subscript
the
indicator.
Warming
scenario.
M
With
the
above
framework,
we
can
get
a
proxy
of
a
symmetry
which
we
call
p-r-e-f
p-I-f-e-t
is
calculated
as
the
annual
e-t
change
under
warm
season.
Warming
divided
by
annual
interchange
under
cool
system
and
pif
q
is
calculated
similarly
using
stream
flow.
If
pif
is
lighter
than
one,
it
means
other
response
happens
under
warm
season
working
if
it
is
smaller
than
one,
it
means
a
larger
response
happens
under
crucial
volume.
M
If
pif
is
smaller
than
zero,
it
means
the
response
is
under
the
two
worming
scenarios
in
our
experiment
to
show
the
different
opposite
signs
in
stream
flow
response.
Based
on
the
long-term
virtual
balance,
I
would
expect
the
pif
et
to
be
close
to
pif
q.
In
any
event-
and
this
is
confirmed
by
this
bar
cloud-
the
first
row
is
prfet
and
the
second
row
is
pif
q.
We
can
see
that
they
are
very
similar.
M
This
ensures
that
we
can
shift
our
focus
to
the
variation
of
pif
et,
because
the
annual
e
t
change
is
the
combined
effect
of
both
water
and
energy
supply
change.
I
want
to
identify
if
there
is
the
season
that
dominates
the
annual
util
change.
This
is
a
plow
showing
the
itchy
response
to
seasonal
warming.
The
first
row
is
each
response
to
warm
season.
Warming
and
the
second
row
is
each
response
to
the
cruising
women.
You
can
see
that
under
warm
season,
warming
annual
heat
change
is
dominated
by
warm
season,
interchange
and
under
cruises,
and
warming.
M
The
annual
eating
change
is
dominated
by
cool
season
interchange.
Both
two
seasonal
et
changes
are
mostly
caused
by
the
season's
temperature
warming
instead
of
water,
because
the
water
availability
at
the
beginning
of
those
swarming
seasons
hardly
changes
in
our
experiments.
So
we
hypothesized
that
the
relative
rank
of
p-I-f-e-t
across
basins
can
be
approximated
by
the
ratio
of
worm
season,
each
to
temperature
sensitivity
to
cruises
and
easy
to
temperature
sensitivity
to
test
this
hypothesis.
We
compare
the
pif
values
under
the
ett
sensitivity
ratio
here
across
the
four
basins.
M
Using
multimodal
mean
we
can
see
that
the
three
proxies
share
very
similar
relative
ranks.
They
also
confirmed
it
across
the
western
u.s
at
hackathe
level,
not
only
in
terms
of
the
relative
rank,
but
also
in
the
absolute
magnitude.
They
are
also
similar,
suggesting
that
the
relationship
is
quite
robust
across
different
hydroclimatic
conditions.
M
Therefore,
if
we
find
what
controls
the
basin's
seasonal
ichito
temperature
sensitivities,
we
can
explain
why
the
pii
of
enhanced
pif
q
differs
across
basin.
To
do
this,
we
choose
a
series
of
variables
related
to
et
and
selected
two
most
representative
variables
here,
namely
temperature
and
the
gi
w.
The
gi
w
is
gross
incoming
water
defined
as
a
student's
precipitation,
plus
the
soil,
water
and
snow
water
storage
at
the
beginning
of
the
season.
The
red
points
denote
the
warm
season
value
under
warm
season
warming
for
each
year,
and
the
blue
points
are
for
cruises.
M
The
previous
slide
shows
the
relationship
within
each
basis
across
different
years,
and
this
slide
compares
the
relationship
between
the
variables
across
different
bases
under
cooler
temperature
in
the
cool
season.
The
e2
temperature
sensitivity
increases
at
basins
with
higher
temperature
well
under
a
warmer
temperature.
The
two
variables
show
different,
very
opposite
variation.
M
Another
look
at
this
relationship
across
the
entire
western
uos,
as
also
shows
the
first
increasing
than
decreasing
pattern.
This
pattern
has
an
important
indication,
for
example,
in
the
red
figure.
The
blue
points
represent
a
cooler
basin
and
the
red
points
denotes
a
warmer
basin.
The
cooler
basin
has
generally
cooler
temperature,
so
it
has
higher
82
temperature
sensitivity
in
the
warm
season
and
low
rated
to
temperature
sensitivity
in
a
cool
season.
Therefore,
the
one
two
cruises
and
82
temperature
sensitivity
ratio
for
the
cooler
basin
is
larger,
since
this
ratio
is
describing
the
response
of
symmetry.
M
So
a
higher
ratio
also
means
that
the
cooler
basin
responses
have
a
stronger
preference
for
the
warm
season.
Vermin.
Such
crew
basins
are
exactly
the
columbia
and
the
colorado
amount
of
force.
Answering
the
second
question
before
I
explain
that
into
temperature
sensitivity.
Thank
you
variation
with
temperature
by
attributing
the
seasonality,
change
to
failed
processes
using
the
permanent
his
framework
and
showing
the
red
figure.
The
red
curves
is
model
simulated
et
change,
and
the
brown
curve
is
the
sum
of
the
foul
individual
processes
contribution.
M
We
can
see
the
increasing
part
of
et2
transmitivity
to
temperature
is
mostly
contributed
by
the
pink
and
the
purple
curves,
which
are
reflecting
enhance
the
rate
of
water,
vapor,
holding
capacity
increase
and
the
decreasing
part
is
due
to
many,
the
blue
and
xeon
curves.
One
is
the
increase
in
surface
resistance
and
a
higher
vapor
pressure
deficit.
Another
is
the
snow
melt,
albedo
feedback
which
causes
less
radiation
increase
when
there
remains
less
low
pack.
Some
other
complex
reasons.
Please
see
our
paper
at
the
end
of
presentation.
M
We
also
discuss
the
effect
of
radiation
on
heat
temperature
sensitivity.
We
only
show
the
main
conclusion
here.
Due
to
time
limit
and
acoustic
warming,
the
warmer
basin
tends
to
melt
more
snow
and
has
lighter
orbital
decrease
and
radiation
increase,
leading
to
a
larger
ec2
temperature
sensitivity
under
warm
seasonal
warming,
a
warmer
basin
generally
have
less,
no
pack
remain
so
less
albedo
decrease
occur,
underwarming,
leading
to
a
smaller
issue
temperature
sensitivity
than
a
cooler
basin.
These
results
are
consistent
with
our
financial
previous,
several
slides.
M
In
summary,
we
find
cooler
basins
tend
to
have
a
larger
annual
rent
of
response
to
warm
season
relative
to
crucial
warming.
This
explains
why
the
colorado
and
columbia
rivers
and
has
a
stronger
stream
flow
decrease
under
warm
season.
Volume
is
partly
explained
by
the
variation
of
it
to
temperature
sensitivity
in
response
to
a
compound
set
of
lens
and
processes,
and
it
has
implications
for
better
understanding
the
future
streamflow
change
under
asymmetrical
warming.
Future
here
are
my
selected
references
and
that's
all
for
details.
M
P
Yeah
I
wanted
to
both
applaud
and
raise
my
hand
yeah,
very
nice
presentation.
Thank
you
yeah,
it's
really,
it's
really
convincing.
I
think
the
only
question
I
had
like
do
you
had
these
sensitivities.
I
I
know
a
little
bit
about
them
from
trying
to
estimate
them
from
observations,
and
it's
very
it
seems
very
difficult.
Do
you
do
you
feel,
like
the
models
are
more
or
less
in
the
right,
ballpark
and
like
how?
How
would
you
conclude
that.
P
Yeah
basically
yeah
like
what
what's
the
what's,
the
real
sensitivity
like
from
the
real
world.
M
The
model
yeah
I'm
using
the
model
in
this
in
this
paper,
and
we
can
see
they.
This
model
has
some
an
uncertainty
in
their
absolute
magnitude
and
the
thing
for
my
work
is:
those
modes
are
calibrated
in
stream
flow
and
the
precipitation.
I
come
from
the
observations,
so
the
swim
simulation
is,
is
mostly
controlling
the
annual
transportation,
I
believe,
and
in
terms
of
whether
it's
difficult
to
to
calculate
the
evap
transformation
to
temperature
sensitivity.
M
I
think
it's
more
difficult
in
in
observation
than
in
motor,
because
if
we
use
remote
sensing
product
of
version
from
transparent-
and
there
are
some
water
balance-
closing
closure
budget
error,
and
it
is
also
hard
to
explain
whether
the
generation
measurement
is
correct
or
not
the
only
thing
we
can.
M
We
can
kind
of
use,
maybe
the
flux
towers,
that
can
measure
the
electrons
variation
and
we
can
use
the
temperature
measurement
too,
and
the
precipitation
and
so
on,
but
I'm
not
quite
familiar
in
that
field,
but
to
constrain
the
variability.
I
believe
there
are
many
methods
that
can
combine
with
different
models.
For
example,
the
observation
constraint
sensitivity
analysis
if
we
can
find
direction
relation
to
temperature
sensitivity
as
having
a
good
linear
regression
relationship
with
another
variable.
That
is
observable
in
the
different
observations
and
this
relationship
is
maintained
across
different
alexia
mp5
models.
B
You,
okay
thanks,
so
we
just
have
10
minutes
left
of
the
actual
session.
So
we'll
use
this
for
some
discussion
now.
Okay,
so
we
have.
We
can
really
talk
about
anything
now.
I
won't
go
through
kind
of
an
update
because
we
already
did
that
on
monday,
but
I
have
these
discussion
points.
B
But
we
can
talk
about
anything
if
anyone
wants
to
say
anything.
So
some
questions,
though,
that
we
thought
might
be
interesting
to
discuss
is
the
usual
one
of
what
simulations
would
you
like
to
see
performed
by
the
working
group
at
this
point,
we're
kind
of
in
the
middle
of
the
allocation
cycle?
So
we
don't
need
to
write
our
proposal
this
year,
so
next
year
will
really
be
when
we'll
need
to
think
of
new
plans,
but
it's
always
good
to
keep
ideas
coming
in
case
we
have
extra
compute
and
time.
J
B
Up
one
question:
we
thought
maybe
needs
to
be
addressed
because
a
number
of
people-
I
guess-
have
asked
me
about
this-
and
I
know
sarah
as
well
as
we
currently
don't-
have
a
slab
ocean
that
is
supported
for
csmt
and
the
working
group
is
going
to
perform
instead
of
the
slab
ocean
pre-industrial
control,
where
the
plan
is
as
young
o'quan
is
running
a
pre-industrial
control
with
this
new
pencil
model,
which
is
basically
a
mixed
layer,
physics
at
every
grid
point.
B
B
B
B
A
This
is
angie,
I
was
just
gonna.
I
was
just
saying
in
the
chat
that
I'm
using
some
aqua
planet
slab
ocean
version
of
something
similar
to
cesm2.
It
might
be
cam
six
beta
the
beta
release
of
something
that
I
got
from
brian
medeiros,
but
I'm
using
that
for
these
itcz
map
simulations,
and
I
think
it's
really
valuable-
to
have
this
level
ocean
configuration.
B
Yeah,
I
think
it's
the
the
getting
the
q
flexes
for
the
more
realistic
version.
I
know
that
cecile
knows
how
to
do
it,
and
I
know
that
it's
not
totally
straightforward.
B
I
don't
think
she's
here
because
she
was
in
the
other
session,
but
it
sounds
like
something
that
could
be
done
and
provided
if
there
was
a
demand
for
it
for
the
kind
of
real
world
situation.
I.
A
A
Yeah
I
mean
I
think
that
that
should
be
feasible
for
realistic
configurations
too.
I
don't
know.
B
C
I'll
add
about
the
aquaplant
and
computing
key
fluxes.
I've
done
this
for
paleo
configurations
and
there
is
supported
tools
for
that
yeah.
I
agree
with
angie
that
I
it's
not.
I
mean
it's.
It
takes
a
little
work,
but
it's
definitely
something
for
people
who
want
to
use
it.
It's
a
valuable
tool
to
keep
around
and
the
tools,
I
think,
are
there.
If
they're
not
perfect
for
season
two,
you
know
I
don't
see
why
they
could
not
be
made.
You
know,
adjust
yeah,
you
know,
with
software
engineering
help.
A
B
Well,
I
think
this
is
yeah.
Resources
are
already
dedicated
to
the
pencil
model.
I
guess
this
is.
This
would
maybe
be
me
talking
to
cecile
and
trying
to
write
a
website
that
would
yeah
set
up,
but
so,
if
there's,
if
there's
a
lot
of
interest
in
it,
maybe
yeah
it
wouldn't
be
too
difficult.
B
So
I
don't
know
if
young
no
is
on
this,
but
I
think
last.
I
heard
I
emailed
him
a
couple
of
weeks
ago
and
they
were
just
kind
of
seeing
if
the
pre-industrial
control
was
stable
and
so
it's
he
signed
a
positive
that
he
would
use
up
the
computing
time
that
was
allocated
before
the
fall.
So
that
would
be
for
the
pre-industrial
control.
B
And
then
I
don't
know
if
they
have
plans
to
write
a
paper
first,
but
I
think
it
probably
expects
sometime
later
this
year
or
early
next
year
will
be
my
guess
if
everything
goes
to
plan
okay,
so
it
seems
like
no
one
does
not
want
a
slab
ocean
model
and
then
71
percent
do
so
that's
pretty
clear.
B
A
F
B
Kind
of
it's:
if
you
want
it,
you
can
definitely
get
it
so
yeah
patrick
callahan
has
implemented
it.
It
hasn't
been
engineered
into
a
release
branch
fully
at
this
point
and
the
software
engineer
who
is
going
to
do
that
is
busy
with
other
things.
So,
if
you
want
to
use
it,
you
can
get
in
touch
with
me
and
I
can
find
out
what
version
on
github.
You
need
to
download
to
be
able
to
use
it,
but
it
should
be
coming
at
some
point:
it's
not
in
csm2.
At
this
point.