►
From YouTube: Alan Wallcraft 2020 05 11
Description
No description was provided for this meeting.
If this is YOUR meeting, an easy way to fix this is to add a description to your video, wherever mtngs.io found it (probably YouTube).
A
A
So
icon
was
the
first
widely
used
al-based
ocean
model,
but
it's
20
years
old
and
lots
of
the
code
is
still
from
I
need
an
older
Micon
code
base.
A
So
for
the
test
case.
It's
a
10-year
simulation
using
the
Goss
3.1
configuration
but
without
icon
specific
features.
41
hybrid
layers
in
the
vertical,
with
one
meter
top
layer
in
high
con.
The
fixed
vertical
coordinate
is
z,
Sigma
Z,
which
I'll
explain
later
and
in
the
test
case
it's
the
star
only
and
in
high
com.
A
We
can
have
different
Target
densities,
for
example,
say
in
the
Black
Sea,
but
in
this
case
later
the
same
Target
density,
I
suppose
everywhere
we
use
a
kppmx
layer
and
we
repeat
one
typical
year
of
cfsr
for
10
model
years
using
the
larger
neweggular
bulk
stress
parameterization,
with
absolute
winds.
A
Some
of
the
differences
between
haikum
and
mom
icon
is
non-bus
and
ask
the
equation
of
State
in
icon.
For
Sigma
2
is
a
17
term
rational
function
and
in
mom
6
we're
using
the
right
pressure.
Separable
equation
of
State
for
efficiency.
A
High
Comm
is
incompressible
with
with
a
theoristic
correction
versus
compressible
Imam,
the
isopectinal
regrading
in
in
icon,
we're
using
layer
averages
and
in
mom
six
we're
using
layer
interfaces
as
the
target
Icom
no
slip
mile
six
free
slip
and
size
4
versus
assist
two
for
the
the
sea
ice
model
and,
generally
speaking,
the
parameters
are
tuned
for
high
com
and
may
not
be
optimal.
For
month.
Six.
A
So
the
10
years
we
were
just
repeated
and
Atmospheric
forcing
the
snapshots
here
are
in
in
logical
array
space.
So
you
can
see
the
tripod
grid
with
the
Arctic
bipolar
patch,
starting
at
47
North,
and
it's
hard
to
see
details
really
in
a
global
plot.
But
you
can
see
these
are
broadly
similar,
similar
Solutions
after
10
years,
both
were
run
on
a
query.
Xd4
on
1800
cores
mom
runs
1.5
times
faster
than
icon.
A
A
So
there's
been
several
presentations
about
what
is
ale
in
in
the
previous
webinars.
A
The
basic
idea
is
we
solve
the
layer
continuing
equation,
so
we
move
the
layers
and,
and
then
we
remap
them
back
to
the
desired
layer
structure.
The
fluid
does
not
move
at
the
layer.
A
The
advantage
of
that
approach
is
that
the
unique
mapping
from
a
monotonic
profile
to
the
Target
entities
in
the
case
of
Icon,
you
can
have
several
different
regrightings,
which
give
you
the
same
layer,
Target
densities,
which
can
cause
problems.
If
you
don't
choose
the
right
one.
A
This
is
a
little
bit
of
a
an
aside,
because
in
the
test
cases
we
don't
use
the
sigma
Z,
but
this
is
showing
the
kind
of
thing
you
can
do
with
an
L
coordinate
in
the
earlier
graphs
3.0
configuration
we
had
32
layers,
one
of
the
things
here
is:
you
know
when
you're
running
operationally
every
day
it's
got
to
run
fast
enough,
and
so,
and
so
we
tended
not
to
add
layers
unless
we
really
thought
it
felt
like
we
had
to
so.
A
A
The
problem
is,
you
know
that
you
then
don't
have
any
structure
in
in
the
velocity,
for
example,
and
and
you
know
that
was
causing
some
issues
for
us
so
in
in
ghost
3.0.
A
You
know
we
added
nine
layers
in
it
and
then
they
were
all
near
the
surface
and
so
14
and
14
of
those
like
14
layers
now
are
terrain
following
and
the
schematic
shows
you
how
that
works.
Essentially,
the
the
the
figures
are
bottomed
out
on
the
x-axis
and
interface
depth,
on
the
y-axis
and
in
deep
water.
Essentially,
the
fixed
coordinates
are:
are
horizontal,
the
the
geopotential.
A
A
For
the
mom
sex
hikon
twins,
essentially,
we
just
use
the
bit
that's
deep
water
from
there.
That
is,
you
know
we
at
96
meters
in
Shallow.
We
continue
to
use
the
coordinates.
We
don't
go.
We
don't
switch
to
Rainbow
line
and
in
high
com,
there's
a
separate
parameter
that
can
set
a
minimum
depth
at
which
you
switch
device
in
in
months
six
and
in
icon
you
can.
You
can
just
do
that
by
setting
the
target
density
very
light.
A
Steps-
and
you
can
see
that
in
the
Southern
Ocean
in
the
top
thousand
meters,
we
are
always
in
fixed,
coordinates
all
the
way
down
to
a
thousand
meters
and
even
in
Middle
attitudes,
we're
we're
deeper
than
96
meters
when
we're
in
we're.
In
always,
the
coordinates
one
take
away
from
this
is
in
fact,
even
though
we're
using
a
different
approach
for
Target
densities.
A
These
are
very
similar
kind
of
looking
sections
one
it
one
issue
with
the
algorithm
we're
using
for
and
try
Martin
d-trainment
is
is
that
you
can
get
thick
thin,
thick
thin
layers
in
icon,
and
if
you
look
at
40
North,
you
can
see
that's
happening
here
now
remember
this
is
a
three-year
mean,
so
that
means
those
layers
are
attending.
This
plot
are
always
then,
and
so
essentially
degenerated.
We
effectively
have
less
in
that
area.
We
have
less
degrees
of
Freedom
as
we
than
we
think
we
had.
A
A
So
this
is
comparing
High
command
to
Global
Drifter
Eddie,
Canada,
energy,
the
lowest
two
plots
are
the
same
plot
and
therefore
Drifter
eke
over
30
years
in
the
top
two
panels
are
three
year
means
from
the
moles
and
again
you
can
see
in
Broad
outline
that
the
strong
currents
are
represented.
D
A
The
Drifters,
that's
about
14
of
the
global
area
between
67
60
North
in
haikum,
it's
about
19
percent
and
Mom
is
26
percent,
so
away
from
strong
currents.
Mom
is
giving
us
less
eke
because
The
Drifter
EKU
is
30-year
average.
So
it's
got
lots
of
information
availability.
It's
hard
to
know
what
the
right
answer
there
is,
whether
you.
C
E
A
A
So
in
this
case
layer
15
is
representing
potentially
just
this
little
piece
of
the
of
the
West
Pacific
everywhere
else.
Essentially,
it's
in
Z
coordinates
High
Congress,.
A
A
Mom
fixes
this
problem
by
integrating
the
finite
volume
pressure
gradient
using
a
numerical
quadrature
and
a
simple
way
to
look
at
the
hemoristic
instability
has
to
plot
the
difference
between
snapshots
12
hours
apart
and
what
you
see
on
the
left
here
is:
is
the
service
editor?
What's
almost
certainly
the
residence
ability
in
in
high
com
and
not
at
all
in
mom
sex,
which
you
know
would
be
expected
from
the
theory?
But
it's
nice
to
know
that?
That's
that's
true.
In
practice,.
A
A
A
A
Now
the
black
bar
on
the
icon
plot
at
the
lower
left,
is
a
section
that
I'm
plowing
in
the
next
slide.
A
This
is
probably
the
largest
difference
between
the
68
seconds
we
routinely
make
the
other
one
that
had
a
large
difference
was
the
Air
Patrol
on
the
current
I
showed
earlier.
A
So
this
is
a
crosstalk
section,
as
I
showed
you
mom
six
has
lost,
left
38.
and
is
further
from
climatology
than
than
icon
and
there's
almost
no
Solange
variation
here
in
Mom's
works.
A
So
one
of
the
things
we
care
about
when
we're
doing
these
high
resolution
runs
of
the
western
Memory
Gardens.
A
When
we
do
assimilation,
the
assimilation
can
correct
for
some
deficiencies
in
location
of
these
western
boundary
currents,
but
if
your
model
doesn't
represent
them
accurately
as
soon
as
you
go
into
forecast
mode
beyond
the
the
assimilation
it
starts.
Reverting
back
to
you
know
the
wrong
state
in
some
sense,
so.
A
There
is
no
variation
in
the
location
of
the
Gulf
Stream
and
Icom
is
consistent
with
that.
Icon
is
separating
at
Cape
Hatteras
as
it
should,
although
it's
slightly
too
far
south
as
it
enters
the
the
rest
of
the
domain.
A
A
A
A
Been
easier
to
illustrate
with
a
movie
but
I
wasn't
sure
that
was
going
to
work
a
problem
we
have
with
this
simulation,
there's
mom
sex
Loop
current.
It
is
never
migrate
into
the
Western
Gulf.
A
Looking
at
the
snapshot,
which
happened
to
be
the
same
day,
there's
no
reason
why
they
should
be
the
same.
It's
a
simulation,
but
you
can
see
that
in
high
com
and
Eddie
has
detached
from
the
low
current
and
there's
a
smaller
and
there's
a
Neti.
That
is
still
there,
which
is
migrated
into
the
Eastern
golf
in
the
mom
sex
case.
The
the
the
loop
current
configuration
is
a
perfectly
reasonable
configuration,
but
in
fact
we
never
get
detached
fully
detached
Luke
guard
edits.
What
happens?
A
And
if
you
look
at
the
secretary
variability
over
three
years,
you
can
see
that
and
in
high
com.
You
can
see
the
past
that
those
Luke
guard
editors
take
into
the
Eastern
Gulf
and
in
the
monks
this
case,
there's
more
variability
concentrate
in
one
area,
but
you
never.
It
never
gets
into
the
Eastern
golf.
A
So,
in
summary,
I
said
that
mom,
sex
and
Hiking
are
very
similar
models,
but
of
course,
mom
was
initially
developed
for
climate
applications,
so
so
how
well
does
it
do
at
the
higher
resolution
and
maybe
for
Global
Ocean,
Prediction
and
I
would
say
that
the
results
are
broadly
similar.
A
A
A
In
question
two
we
haven't:
we
now
have
an
Atlantic,
only
configuration
which
is
a
bit
cheaper
to
run,
which
we're
using
to
explore
some
of
these
issues.
But,
generally
speaking,
mom
is
still
a
good
candidate
for
future
high
resolution.
Protection
Systems
and
one
of
its
advantages
is:
it
allows
the
same
model
to
use
for
the
synoptic
and
climate
time
scales,
I,
think
here
by
climate
I'm,
probably
meaning
climate
in
the
in
the
nwp
sense.
A
So
in
other
words,
it's
a
simple
location
for
for
an
operational
Center
to
have
one
model
rather
than
two
icon.
3
is
mine
compared
to
be
high
gun
like
and
using
icon
tools
and
I
think
long
term.
There
probably
won't
be
a
icon
3.
That
is,
it
really
is
Mom
six,
but
it
does
simplify
the
transition
from
Mom
6
for
existing
users.
B
D
Alan
Chris,
okay,
Enrique
I,
had
a
question.
B
Yeah
Alan
quick
question
when
you
use
the
terrain
following
coordinates:
does
that
put
a
constraint
on
the
gradient
of
your
bathymetry.
A
No
one
of
the
advantages
of
using
Sigma
Z
is:
is
that
you're
only
you're
only
terrain
following
in
shallow
water,
so
you're
not
doing
the
continental
slope
which
tends
to
have
you
know
larger
gradients,
but,
generally
speaking,
we
don't
try
and
make
the
we
don't.
Try
and
smooth
the
bitometeries
to
to
meet
some
of
the
criteria
you
might
use
if
you
were
using
a
conventional
terrain,
fly
model
and
conceivably,
that's
a
problem,
but
you
know
we're
on
we're
on
the
Shelf
we're
doing
a
lot
of
mixing
anyway.
You
know.
B
E
Alan
on
page
on
slide
four,
if
you
could
go
there,
there's
some
terminology
I'm
a
little
confused
about
I
just
want
to
get
clarified
on
on
the
use
of
terminology.
I
believe
you
said
that
the
high
Comm
is
non-boozonesque,
but
that
you
use
a
in
incompressible
with
thermobilicity
versus
compressible.
I
was
I,
mean
incompressible
versus
non-businesque
numbers
and
ask:
is
it
income?
Is
a
compressible
fluid?
Typically,
that's
the
the
terminology.
People
usually
use.
So
could
you
clarify
that
point?
What
you're
talking
about
there.
A
So
so
I'm
using
booze
next
in
the
sense
of
you
know
what's
concerned,
is
it
layer,
volume
or
layer,
Mass
and.
A
E
E
E
E
A
A
So
we
have
both
by
harmonic
and
laplacian.
Okay.
A
F
Yeah,
the
I
think
you
indicated
that
you
were
going
to
use
the
Atlantic
configuration
to
iron
out
some
of
the
issues.
I
think
that
configuration
will
be
okay
for
the
Gulf
of
Mexico
issues,
but
the
any
information
you
get
from
the
Atlantic
for
a
mark
and
its
vertical
extent
or
whatever
will
be
highly
dependent
upon
what
kind
of
boundary
conditions
you
use
on
the
Northern
side
and
that
may
not
be
necessarily
translatable
to
the
full
Global
configuration.
A
F
A
Mean
we've
certainly
seen
already
that
because
in
fact
we're
you
know
we're
imposing
climatology
on
the
Northern
boundary
it.
It
isn't
exactly
the
same
as
the
global
model
and
when
you
look
at
the
overturning
circulation,
but
it's
but
in
the
same
configuration
as
the
global
model.
It's
it's
somewhat
similar,
and
so
it
gives
us
at
least
a
starting
point
for
looking
at.
You
know
what
what
could
we
change
to
make
it
different.
D
Okay,
Dan
Witt.
C
D
C
See
that
this
is
a
point
of
clarification,
I
see
that
you're
using
the
large
or
bulk
plus
parameterization
with
absolute
wins
and
I,
presume
that
that
means
you
can
clarify
America
wrong
but
you're
not
considering
the
ocean
currents.
When
you
calculate
the
stress
and
I
guess:
I'm
wondering
if
you've,
why
you're
doing
that
and
if
you've
tried
the
relative
currents
or
some
variant
of
the
relative
winds
with
with
the
ocean
currents
and
found
that
that
has
deficiencies
in
your
case.
A
So
we're
using
so
when
we
run
icon
will
use
the
core
three
algorithm,
not
not
unilateral,
generally
speaking,
but
because
mom
actually
has
laundry
like
a
built
in.
We
switched
it
using
that
and
generally,
when
we
run
high
com
for
prediction,
we
would
use
a
relative
winds.
A
It
tends
to
dissipate
that
is
and,
and
so,
and
so
they
tend
to
last
list
less
time
and
and
ocean
currents
are
generally
less
strong
and.
A
In
theory,
that's
that's
correct
that
that
is,
you
know,
but
the
problem
is,
you
know:
we're
trying
to
run
at
essentially
a
sub-optimal
grid
resolution
and
yet
still
get
good
answers,
and
so
sometimes
you
go.
You
go
with.
B
A
Monument
I.