►
From YouTube: CTSM Tutorial - Wednesday
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
Introduction
to
what
we're
going
to
be
doing
here
and
then
kind
of
launch
into
it,
so
first
off,
I'm
super
excited
that
so
many
of
you
joined
this
kind
of
feedback
or
this
kind
of
participation.
That's
really
critical
for
a
for
a
community
model
like
ctsm,
and
so
thank
you
for
for
joining
us,
and
I
hope
the
next
couple
days
are
productive
and
and
a
good
learning
experience
for
everyone.
A
A
Consider
the
new
ideas
that
we're
sharing
with
each
other
show
appreciation,
encourage
innovation
and
then
acknowledge
teamwork,
and
I
think
kind
of
following
these
principles
helps
us
will
help
us
have
a
you
know,
a
productive,
collegial
meeting
and
to
start
with,
I
want
to
begin
with
a
bunch
of
thank
yous,
so
none
of
this
would
have
happened
without
a
ton
of
work
from
adriana
nagin,
danica
and
me
who
put
the
tutorial
together
and
then
brian,
who
kind
of
behind
the
scenes.
A
Jackie
paulie
and
keith
did
a
ton
of
testing
on
the
tutorial
materials
before
we
passed
them
all
out
to
you
guys,
and
then
software
engineers,
eric
luzak
and
bill
sex,
made
some
kind
of
last-minute
critical
tags
that
were
needed
for
running
these
simulations
on
aws,
elizabeth
and
ryan
have
done
a
bunch
with
tutorial
registration
logistics
web
pages.
A
I
see
that
tracy
and
todd
are
also
here
to
help
kind
of
navigate
for
today,
or
maybe
it's
not
tracy,
sorry
roberta
and
then
also
the
tss
staff.
The
terrestrial
science
fiction
staff
at
ncar
are
going
to
help
with
the
the
practical
session
of
the
tutorial
today
and
also
provide
some
extra
materials
tomorrow
so
and
cars,
and
all
of
this
is
supported
by
the
national
science
foundation,
with
a
big
cooperative
agreement
as
well
as
a
few
other
smaller
project
awards
to
several
of
us
on
this.
A
On
this
slide
danica,
do
you
mind
putting
dropping
these
links
in
the
chat?
These
will
be
helpful
kind
of
for
all
of
us
throughout
the
day,
and
so
one
is
the
survey.
A
I
think
51
of
us
have
done
the
survey
already,
which
is
awesome,
so,
if
you're
participating
in
the
workshop
filling
out,
the
survey
would
be
helpful,
and
this
is
specifically
because
nsf
asked
for
some
ask
for
some
feedback
about
who's
taking
the
tutorial
and
what
what
you
all
are
learning
so
having
this
pre-tutorial
survey
will
be
really
helpful
on
this
event,
page
that
should
have
come
out,
but
it's
on
this
website
and
I'm
gonna
just
go
to
it
real
quick.
A
You
can
look
at
the
agenda
which
has
been
changed
from
what's
posted
on
here.
That's
fine,
but
there's
a
number
of
resources,
and
this
is
kind
of
a
helpful
landing
page
to
get
back
to
so
there
will
be
there's
the
video
that
adriana
recorded.
That's
that's
up!
There's
the
tutorial
materials,
the
readme
that
you
all
started
on
to
get
to
kind
of
get
start
working
on
the
day,
zero
homework,
which
thank
you
for
many
of
you
for
doing
that.
The
lecture
materials
are
up
in
this
google
drive
folder.
A
You
know.
We've
only
got
a
couple
of
these
up
so
far,
but
this
welcome
intro
that
we're
looking
at
now
is
up
here
and
and
dave's
overview.
Talk
is
in
here
now,
so
there
will
be
more
stuff
coming
up
here.
A
Go
back
to
this
and
then
also
there's
a
link
that
danica
put
in
the
chat
for
cesm
lab,
so
brian's
going
to
be
adding
some
more
nodes
to
that
as
dave's
lecturing,
so
that
we
have
more
computing
resources
to
actually
do
the
day.
One
tutorial
today,
but
there's
a
link
to
this,
the
aws
cloud.
A
If
you
need
it
or-
and
you
will
need
it
so,
land
model
tutorials
at
ncar
have
a
pretty
long
history.
This
is
kind
of
fun
to
put
together.
We
started
in
2014
with
a
four
and
a
half
day
tutorial
introducing
clm
4.5.
A
There
were
four
practical
sessions
and
15
lectures
and
all
of
the
practicals
were
done
on
yellowstone.
It
expanded
a
little
bit
in
2016
to
have
22
lectures
and
again
four
and
a
half
days
long
2019
was
our
last
in-person
tutorial.
A
It
was
pretty
much
a
full
five
days,
there's
basically
five
practical
sessions
and
20
lectures
that
students
did
on
on
cheyenne
the
super
computer
that
that
we
have
now-
and
so
this
is
a
big
shift
for
us
and
that's
kind
of
what
I
want
to
emphasize
here.
We've
never
done
it.
Virtually
you
can
see
here.
The
photo
below
is
from
the
2019
tutorial.
A
It
almost
always
snows
in
boulder
around
the
tutorial
in
true
to
forum.
It
snowed
on
on
saturday,
friday
and
saturday
this
week,
and
because
it's
virtual,
we
decided
that
kind
of
given
zoom
fatigue.
We
would
keep
it
short
short
and
sweet,
so
we're
only
doing
two
half
days,
there's
two
practical
sessions
and
two
effectively
two
lectures.
A
So
this
is
very
condensed
and-
and
I
appreciate
that
that
many
of
you
are
going
to
want
more
after
this,
but
because
we're
using
cesm
lab
and
we're
doing
all
of
this
in
the
cloud.
The
idea
is
that
kind
of
moving
forward.
Hopefully
we
can
offer
you
know,
we
don't
need
to
have
this
two
to
three
year,
latency
to
offer
new
tutorials
and
that
we
can
hopefully
kind
of
more
regularly
offer
tutorials.
A
That
that
maybe
are
shorter
and
more
modular
and
that
we
can
build
on
so
we
kind
of
are
hoping
to
build
this
library
of
tutorial
materials
that
use
cesm
lab
in
the
cloud
and
and
be
able
to
to
to
do
this
in
a
variety
of
settings.
So
your
feedback
moving
forward-
I
guess
it
will
be-
will
be
really
really
critical.
A
You
know
one
bummer
is
that
we
don't
get
to
see
each
other
face
to
face,
and
so
hopefully
we'll
get
to
spend
some
time
getting
to
know
smaller
groups
of
people,
but
I
appreciate
that
it
will
have
a
different
flavor
than
past
tutorials
have
so
in
the
in
registration.
So
far,
you
know
we're
kind
of
I
think,
hitting
the
demographic
that
we
want
to
in
terms
of
this
tutorial.
A
Nearly
all
of
you
are
our
students
or
early
career
scientists,
so
they're
60
people
registered
on
three
different
continents
and
and
also
props
to
the
mid
and
advanced
career
individuals
that
are
here.
It's
awesome
to
have
have
some
more
senior
people
joining
to
see
what
you
know
more
early
career
scientists
are
going
to
be
doing
and
then
from
what
you
want
to
be
able
to
to
learn
from
this
tutorial.
You
know
it's
things
like
learn.
A
The
science
understand,
ctsm,
better
use
the
model
run
simulations,
and
then
I
liked
what
ended
up
on
the
trunk
of
this
word
tree
data
basic
and
want
more
so
it's
kind
of
just
a
smattering
of
what
folks
are
interested
in
and
then
you
had.
You
got
to
rate
yourselves
on
your
familiarity
with
ctsm
and
your
proficiency
in
writing.
A
Models
and
again,
you
know
people
aren't
super
familiar
with
a
lot
of
the
science
and
ctss,
so
that's
great
and
also
maybe
aren't
very,
very
confident
their
abilities
to
run
simulations,
and
so
you
know
for
all
of
this.
What
I
want
you
to
what
what
I'd
like
you
to
to
do
is
to
is
to
really
make
connections
with
other
students
that
are
in
the
group.
Ask
questions
both
of
the
the
learning
the
people
in
the
learning
pods
that
we
put
you
with,
as
well
as
the
staff.
A
That's
here
to
help
out
and
then
also
stay
engaged.
So
beyond
the
tutorial
you
know,
feel
free
to
reach
out
and
and
get
help
where
you
need
to,
because
ttsm
is
a
big
complex
model
and,
and
it
really
does
take
a
community
of
people
to
to
understand
it
and
to
work
to
improve
it.
A
So
briefly,
here's
the
kind
of
adjusted
agenda
that
we
have
for
today
dave
will
give
an
overview
presentation
and
and
then
we'll
spend
about
10
minutes
kind
of
getting
started
with
the
learning,
pods
and
and
kind
of
expectations
for
the
for
the
breakout.
I'm
sorry
for
the
practical
session
put
a
short
break
in
there
and
then
we'll
have
you
know
roughly
an
hour
and
a
half
to
go
ahead
and
work
through
these
two
global
scale.
A
Practical
cases
before
adjourning
around
noon
mountain
time
today,
and
so
I
wanted
to
to
introduce
dave
briefly
dave's
a
a
senior
scientist
to
ncar
who's.
I
don't
know
how
long
you've
been
here
dave
for
a
while
and
they've
ran
the
the
land
model
working
group
from,
I
think
2005
until
2019,
and
now
he's
continuing
on
as
the
lead
of
the
ctsm
effort
he's
in
charge
or
has
his
hands
on
a
number
of
projects:
the
permacross
carbon
network.
A
You
know,
chair
of
lumipip,
the
land
use
model
and
a
comparison
project,
part
of
cmap6
and
aside
from
being
a
very
accomplished
scientist,
it's
also
just
a
joy
to
work
with.
So
it's
my
pleasure
to
introduce
dave
lawrence
to
give
us
an
overview.
Talk.
B
Right
so
I'm
gonna
presuming
you're,
seeing
this
all
fine.
Well,
let
me
know
if
not
yeah,
that
looks
good,
so
I'm
gonna,
I'm
gonna,
give
an
introductory
lecture,
and
you
know,
as
will
pointed
out,
we
you
know
typically,
would
have
had
many
many
lectures
to
try
to
cover
all
the
physics
and
the
ecology
and
the
biochemistry
hydrology
of
the
model,
and
we
just
aren't
going
to
have
time
in
this
mini
tutorial,
and
I'm
not
going
to
try
to
I'm
not
going
to
try
to
do
that
here.
Condense
it
into
into
one
30.
B
So
what
I'm
going
to
really
do
is
just
give
you
a
broad
bird's
eye
overview
of
the
model
and
then
talk
a
little
bit
about
how
we
operate,
how
we
develop
the
model
and
then
kind
of
conclude
with
some
some
thoughts,
maybe
ways
to
sort
of
spur
your
thinking
about
how
you
might
use
model
in
your
research,
because
there's
a
lot
of
capabilities
within
ctsm
that
that
you
know
you're
not
even
going
to
learn
during
this
during
this
next
couple
of
days,
but
are
important
to
be
aware
of
I'd
like
to
start
these
sort
of
general
overview
lectures
with
this.
B
This
question
you
know
land
modeling.
Why
and-
and
this
is
a
photo
of
my
my
wife
and
she's-
an
oceanographer-
and
we
have
these
conversations
in
the
kitchen
and
she
likes
provocative
and
she's
always
saying.
Are
you
sure
this
is
really
necessary?
Oceanographers
believe
that
they're
the
only
important
parts
of
the
climate
system?
So
you
know
we
have
these
fun
conversations,
and
I
said
yes,
of
course,
the
land
is
the
critical
interface.
B
The
reason
is,
land
is
critical
interface
to
which
humanity
affects
and
is
affected
by
adapts
to
and
mitigates
global
environmental
change.
So
the
answer
is
actually
very
straightforward.
Land
is
critical
in
understanding
their
system,
but
if
you
want
to
look
into
a
little
more
detail,
why
do
we
need
a
model
like
ctsm?
B
Well,
one
of
the
topics
is
land
atmosphere,
reactions
and
that
you
know
people
have
been
researching
this
for
many
years.
Although
there's
still
a
lot
of
open
questions,
if
you
have
a
rain
event
and
there's
some
soil,
moisture
and
there's
more
evaporation,
does
that
lead
in
these
coming
days
to
increase
precipitation?
B
There
is
evidence
that
there
is
significant
skill
that
you
get
when
you
conditioned
on
initial
amplitude,
slow
motion
anomalies
in
predictions
out
at
relatively
long
time,
scales
30
to
45
days
related
to
soil,
moisture
and
we've
also
seen
some
evidence
that
plant
state
can
affect
that.
B
There's
questions
about
how
land
atmospheric
company
is
going
to
change
in
a
future
warmer
climate
and
could
that
lead
to
a
increased
land,
difference,
skill
and
then
there's
questions
in
terms
of
this
realm
of
land
atmosphere,
interactions
about
how
land
processes
affect
extremes
and
we've
been
able
to
show
using
ctsm
and
other
models
that
you
know.
Land
processes
are
very
important
in
terms
of
determining
the
magnitude
of
the
heat
extremes.
B
A
B
To
study
water,
you
know
what's
going
on
with
the
water
cycle,
you
know.
Big
thing
is
let
the
land
feedbacks
on
droughts
and
floods.
What
contribution
does
the
land
directly
playing
in
in
extending
increasing
the
amplitude
or
the
frequency
of
trout
and
flood?
There's
processes
like
the
snow,
albedo
and
the
snow,
so
temperature
feedbacks?
B
We
know
that
snow
cover
is
is
decreasing.
You
know
how
is
that
affecting
the
broader
climate
system,
really
important
water
and
food
security?
You
know
more
than
sixth
world
populations
depend
on
water
from
seasonal
snowpacks
and
so
be
able
to
predict
how
that
snowpack
may
change
in
the
future
under
under
changing
climate.
B
C
B
Actually
make
it
harder
to
predict
your
seasonal
water
availability,
important
thing
that
we
study
is
water.
Plant
interactions,
you
know,
plant
water
use,
efficiency
is
likely
to
increase
with
co2,
and
this
affects
you
know
water
availability
on
land
and
also
it's
going
to
affect
plants,
plant
health
and
then
ecosystem
health
and
then
there's
others
who
are
interested
in
streamflow
prediction.
Better.
You
know,
prediction
of
stream
flow
is,
is
critically
important
for
for
water
managers,
especially
as
we
have
started
having
less
water
available.
B
Other
reasons
we
saw
the
study
of
the
land
uses,
models
like
ctsm
is
to
study,
land
use
and
land
cover
change
about.
25
of
the
non-iceland
area
has
undergone
some
sort
of
land
cover
change,
so
a
big
fraction
of
the
world
has
has
experienced
land
cover
change.
What
impact
does
that
have?
On
climate?
B
It's
not
just
land
cover
change.
In
fact,
almost
80
percent
of
the
non-iceland
area
is
under
some
sort
of
land
management.
So
maybe
it's
not
land
cover
change,
but
the
way
we
are
using
that
land
say
through
wood
harvest
or
other
parts
like
that
is
changing,
and
this
these
maps
on
the
right
here
show
that
you
know
the
extent
of
the
of
the
fact
that
humans
have
basically
touched
almost
every
part
of
the
land
surface.
B
We
know
that
regionally,
land
use
and
lanca
retain
can
be
impactful
on
surface
climate
as
greenhouse
gases,
and
I
just
mentioned
you
know
this
idea
of
extremes.
Irrigation
we've
been
able
to
show
with
ctsm
can
actually
mitigate
against
heat
extremes
a
bit
of
an
offset
of
climate
change
feedbacks.
B
Furthermore,
in
terms
of
the
carbon
cycle
about
a
third
of
the
direct
stored,
carbon
emissions
have
come
from
from
land
use,
which
is
you
know,
maybe
more
than
you
might
have
anticipated
present
day.
It's
maybe
more
like
10
percent,
but
over
the
course
of
the
entire
stroke
record
about
a
third
is
coming
from
land
exchange
and
another
important
thing
is
deforestation.
B
When
you
do
deforest
it,
you
lose
the
capacity
to
to
pull
up
as
much
carbon
as
you
would
have
been
able
to
without
deforesting.
So
that
has
another
indirect
carbon
impact
that
we
need
to
understand
and
really,
I
think
this
is
one
of
in
my
mind.
One
of
the
more
important
questions
that
we're
trying
to
address
these
days
is,
you
know:
can
we
use
the
land
to
help
mitigate
against
climate
change
through
processes
like
re
or
affordation
or
or
biofuels,
and.
B
Differences
between
urban
and
real
regions
and
pesticide
stress
and
then
the
final
example
I
want
to
bring
here
is,
is
on
carbon
and
ecology.
The
carbon
nitrous
cycle
interactions
on
the
impact
of
the
long-term
trajectory
of
intestinal
carbon
sinker
are
order.
One
questions
that
we
are
trying
to
trying
to
resolve
with
the
model
like
ctsm
and.
B
Really
large
uncertainty
in
the
projected
land
carbon
sink.
This
is
not
just
like
a
you
know
scientifically
interesting
topic.
This
is
impactful
it's
about
contributing
to
about
a
quarter
of
the
total
uncertainty
or
more
in
in
climate
change
on
the
future.
This
uncertainty
in
the
in
the
land
carbon
we'd
like
to
understand
the
vulnerability
of
ecosystems
to
climate
change
and
other
disturbances,
both
natural
and
human,
and
this
then
also
relates
to
ecosystem
services.
What
services
are
forests
providing
and
will
those
services
be
maintained
going
into
the
future?
B
And
then
you
know
another
topic
is
you
know?
Can
we
do
targeted
ecosystem
management?
You
know
maintaining
forest
health
if
we
can
to
mitigate
against
climate
change.
So,
to
you
know,
answer
all
these
scientific
questions.
You
need
a
land
model
and
a
land
model.
Ideally
it's
coupled
to
a
nurses
to
model
and
to
help
address
these
questions.
We've
been
building
on
these
models
over
the
past
30
years
or
so
30
40
years,
and
so
these
models
started
out
as
as
being
relatively
simple
or
very
quite
simple,
and
when
they
started
coupling
to
the
atmosphere.
B
B
I
was
not
involved
yet,
but
we
scientists
involved
in
land
modeling
started
to
add
features
like
interactive
plant,
canopies,
stomata,
resistance,
soil,
moisture
and
accounting
for
for
surface
headers
in
80
and
getting
out
into
say
the
2000s
incorporating
features
like
carbon
cycle
lakes,
rivers
and
wetlands,
groundwater
and
dynamic
vegetation
processes,
and
now,
as
we
get
out,
you
know
into
into
present
closer
to
present
day
representing
processes
like
land
cover
change
and
urban
environments,
nutrients
crops
and
irrigation
and
lateral
flow,
and
we've
really
evolved
from.
B
You
know
this
idea
that
land
is
a
lower
boundary
condition
to
the
atmosphere
in
the
early
days
of
of
of
climate
modeling
to
the
idea
that
land
is
an
integral
component
of
the
earth
system,
and
this
gets
us
to
where
we
are
today,
roughly
with
with
the
community
landmark
version
five
and
that
this
figure,
which
is
you
know,
way
too
complex
to
to
to
go
over
in
in
a
presentation-
and
I
expect
many
of
you
have
already
seen
this
figure
tries
to
capture
the
range
of
processes
that
we
are
currently
modeling
within
within
clm
and
it
captures
you
know
the
surface
energy
budget,
the
water
cycle
and
biogeochemical
cycles.
B
So
you
know
fluxes
at
that
what
we
call
the
column
level
vertical
fluxes.
In
addition,
we're
trying
to
capture
landscape
level
processes
like
rivers,
that
are,
you
know,
moving
moving
mass
or
energy
through
the
system.
The
idea
of
you
know
the
heterogeneity,
land,
use,
change,
dynamic,
vegetation
processes,
glaciers
and
and
whatnot,
and
how
changes
in
these
through
time
can
affect,
can
affect
climate
and
ecosystem
functioning.
B
So
I'm
not
going
to
go
through
and
try
to
explain
this
each
each
arrow
on
here.
You
know
you
could
give
a
lecture
online,
most
likely
it's
based
on
on
years
of
research,
both
field
scientists,
theoretical
work,
modeling
work
from
simple
models,
complex
models,
but
I
just
want
to
give
a
sense
that
you
know
each
one
of
those
each
one
of
those
arrows
or
elements
boxes
in
that
model
can't
include.
B
It
does
include
a
lot
of
complexity,
and
so
that's
something
that,
as
you
begin
to
work
with
the
model
you're
going
to
slowly
try
to
build
up
your
knowledge
base
to
help.
You
understand
your
sciences
coming
out
of
the
model,
but
here's
just
an
example
that
snow
model
I
like
to
use
it
because
it's
we
all
can
easily
conceptualize
and
think
about
snow.
At
least
those
of
us
who
live
here
in
places
like
colorado,
where
we
experience
it
all
the
time,
but
there's
a
lot
that
goes
into
just
the
snow
model.
B
If
you
say
we
have
a
snow
model,
what
does
that
mean?
We
represent?
It's
got
10
layers,
we
represent
processes
like
accumulation
and
the
fresh
snow
density
is
an
important
factor.
That's
currently
a
factor
of
both
temperature
and
wind
speed.
We
have
to
calculate
processes
like
snow
melts
and
refreezing.
There's
no
age.
You
know
fresh
snow
is
going
to
be
brighter
than
than
older
snow,
and
that
has
a
strong
impact
on
the
albedo.
B
We
need
to
keep
track
of
water
and
energy
transfer
across
layers,
the
snow
compacts
after
it's
fallen,
and
so
we
have
three
different
processes.
I
calculate
that
metamorph
system
due
to
temperature
and
wind
overburden,
basically
just
the
weight
of
the
snow
and
melt
and
freeze
cycles
all
affect
the
snow
density.
B
There's
a
sublimation
process,
there's
aerosol
deposition
on
the
snow,
there's
canopy,
snow
storage
and
unloading
and
related
canopy
snow
radiation.
So
when
snow
is
sticking
on
the
trees,
that
obviously
has
a
different
impact
in
terms
of
its
albedo
than
snow.
That's
fallen
down
to
the
ground
a
few
days
after
a
snow
event,
there's
no
barely
a
vegetation
for
a
short
short
vegetation,
there's
no
cover
fraction.
B
So
these
are
all
processes
that
we
capture
to
some
degree
of
fidelity
within
within
ctsm,
and
even
then
it's
not
everything
we
really
would
want
to
have
in
a
in
an
ideal
model.
It's
missing
processes
like
blowing
snow,
sub,
good
variations,
snow
depth
and
depth.
Four,
which
we
know
are
are
are
important
in
in
certain
parts
of
the
ear
system.
B
You
know
another
way
to
think
about.
Complexity
is
also
I
like
to
use
this
example
about
land
management.
So
you
know
I
just
showed
like
a
process
model
level
complexity,
but
there's
also
sort
of
spatially
explicit
level
complexity
on
sort
of
features
that
we're
trying
to
capture
of
how
humans
are
basically
interacting
with
the
land
surface.
So
among
the
things
we
have
and
you'll
hear
a
little
bit
more
about
this
tomorrow
is
that
it's
a
global
crop
model
with
eight
different
crop
types
and
features
like
planting
grain
filling
harvests.
B
We
include
processes
like
irrigation,
industrial
fertilization.
We
also
represent
wood
harvest
to
account,
for
you
know
how
wood
is
being
taken
off
of
the
land
to
be
used
for,
for
industrial
purposes,
building
houses,
etc.
We
represent
our
urban
environments
and
we
also
represent
their
anthropogenic
fire
ignition
and
suppression,
and
of.
B
That
humans
are
interacting
with
the
system
that
we
can
potentially
add
in
the
future.
So
that's
sort
of
process,
level
topics,
there's
also
this
concept
of
heterogeneity,
which
is
a
big
challenge
in
land
modeling,
land
surface
hydrogenated
in
in
cereal,
and
we
deal
with
through
and
most
other
models
of
this
class
deal
through
a
process.
We
call
subject
tiling,
and
so
we
take
a
grid
cell.
We
split
each
grid
cell
up
into
in
our
case,
right
now
we
have
five
different
land
units,
vegetated
lake,
urban,
glacier
and
crop.
B
B
These
different
plant
functional
types
and
if
that
crops
could
be
treated
either
as
a
as
a
plant,
functional
type
or
as
I'll
show,
the
moment
crop
functional
types.
You
know
the
urban
has
different
features
to
it:
calculating
the
details
of
the
urban
and
the
glacier
model.
We
deal
with
multiple
elevation
classes
in
the
crop
model.
We
have
multiple
crop
types
with
irrigated
un-irrigated
crop
one
and
a
cotton
for
example.
B
So
you
hear
a
little
bit
more
about
this
tomorrow,
but
the
main
point
I
want
to
stress
here
is
this
is
how
we
deal
with
engineering
we
take
from
modis
and
others
observational
data
sets.
We
can
get
the
area
weights
for
each
one
of
these
surface
types,
both
at
present
day
and
historically
changing
through
time,
and
the
errors
here
indicate
the
allowed
transitions
between
these
different
these
different
types.
B
So
we
can
have
lake
area
change,
glacier
area,
change,
change
in
weights
of
plant,
functional
types,
changing
weights
of
of
cropland
types
and
switching
between
cropland
and
vegetated
all
to
account
for
how
the
land
surface
is
changing
either
due
to
human
or
natural
processes,
and
in
fact
the
urban
model
is
now
allowed
to
be
transient
as
well.
B
I'm
going
to
skip
over
this
an
interesting
time.
Okay,
so
you
know
so
we
have
this
model,
that's
great,
because
we've
got
a
lot
of
processes
that
are
all
coupled
together.
You
know
what
does
it
all
mean
we,
so
one
thing
that
we
need
to
do
and
you'll
need
to
do
is
you
know,
develop
methods
to
assess
the
model
I'll
get
to
that
in
a
moment
here.
Sorry
I'll
forget
exactly
one
of
my
slides.
So
I
just
want
to
point
out
here:
ctsm
is:
is
a
community
tool?
B
It's
able
to
use
it
in
many
different
ways.
You
can
couple
it
to
tsm
or
just
couple
the
csm,
and
then
you
can
ask
all
sorts
of
questions
about
land
atmosphere,
reactions,
climate
variability,
air
quality,
climate
change
and
weather.
It's
also
coupled
to
the
wharf
regional
model.
If
you
want
to
look
at
high
resolution
processes-
and
it
can
be
coupled
fairly
straightforwardly
to
other
other-
and
it
is
coupled
to
other
assistant
models
or
the
regional
models,
so
it's
used
in
coupled
mode
a
lot,
but
it's
also
used
and
actually
developed
by.
B
You
know:
communities
who
are
interested
in
in
terrestrial
processes
like
hydrology
or
ecology,
or
biology
chemistry
or
the
cryosphere
or
societal
or
human
dimensions,
and
these
groups
can
you
can
use
the
model
in
land
only
mode,
and
these
are
the
groups
that
often
tend
to
to
work
to
help
improve
the
model
through
time-
and
you
know,
the
model
has
been
improved
steadily
through
time.
Since
I've
been
involved
for
about
20
years,
we've
gone
through
many
versions
of
the
model.
B
Actually,
I
started
with
lithium
3.5
when
I
arrived
here
and
I'm
not
going
to
go
through
all
the
details,
but
at
each
iteration
of
the
model
we
have
been
adding
in
collaboration
with
the
community.
Many
new
features,
so
this
one
is:
has
carbon
nitrogen
permafrost
enabled
transient
light
cover
change?
We
released
clm
4.5
in
in
2013
and
added
a
whole
bunch
of
new
features.
Additional
features
vertical
result
of
carbon
methane
emissions.
B
You
know
fire
processes,
hydrology
clm5,
the
scale
of
change
increased
by
maybe
not
an
order
of
magnitude
by
by
by
quite
a
bit-
and
you
know
now
you
can
you
can
see
how
many
people
are
working
on
it.
Yet
more
than
50
researchers
from
15
different
institutions
and
we're
involved
in
the
development
of
it,
and
you
know
we
had
a
lot
of
different
priorities.
B
You
know
hydrology
is
more
consistent
with
state-of-the-art
understanding,
more
ecologically
relevant
nutrient
carbon
and
water
dynamics,
representing
land
management,
in
a
much
more
detailed
way,
allowing
for
prognostic
greenland
ice
sheet
model
and
many
more
things,
and
we
put
this
all
together
and
there's
a
whole
bunch
of
new
features
within
clm5
and
you'll
be
using
essentially
a
a
version
of
cn5
and
some
of
your
tutorial
stuff.
Today,
we've
now
moved
on,
since
it's
now
a
few
years
old.
We
have
a
bunch
of
new
things.
B
In
addition,
so
essential
challenge
for
for
you
and
for
everyone
is
that
you
know
with
all
these
features
in
the
model.
Is:
are
these
models?
You
know
getting
better
more
realistic,
or
are
they
just
more
complex?
B
Other
questions
we
need
to
address
or
deal
animals
need
to
be
more
complex,
there's
not
just
etsm
there's,
you
know,
there's
many
other
land
models
that
have
been
developed
by
other
groups
around
the
world
and
trying
to
understand
the
divergent
responses
in
all
these
models
is
another
big
aspect
of
the
science
that
we
do
and
all
this,
though
you
know
in
terms
of
this
tutorial,
I
think,
and
the
concept
I'm
going
to
kind
of
go
through
several
times
for
the
remaining.
You
know.
C
B
B
How
do
you
even
assess
whether
or
not
your
model
is
good
enough
and
that's
that's
a
big
part
of
of
the
science
that
goes
on,
and
so
one
of
the
ways
that
we
do,
that
is
through
analysis
packages,
and
so
one
package
that
exists
out
there
is
the
international
landmark
benchmarking
package,
a
project
that
I've
been
involved
with
for
for
many
years
now,
mainly
sponsored
by
the
department
of
energy,
and
this
is
a
land
diagnostics
package
that
integrates
analysis
of
30
plus
variables
against
70,
plus
global
regional
insight
level.
B
Data
sets
with
a
whole
range
of
metrics,
for
things
like
mean
square
bias,
spatial
data,
correlation
internal
variability
function,
relationships
and
on
the
bottom
just
shows
you
know
a
sampling
of
some
of
the
types
of
of
plots
that
you
can
find
within
this
package,
and
so
you
know
what
we
do
then
is
when
we
have
a
new
version
of
the
model,
we
will
run
the
island
package
against
the
old
version
of
the
model
and
see
how
the
model
change
and
did
the
model
get,
did
the
model
get
better,
did
them
all
get
worse,
and
what
you'll
find
typically
is
that
the
model
gets
better
in
some
ways
and
gets
worse
than
others,
and
so
I'm
not
going
to
get
a
chance
here
to
to
go
through
the
island
package.
B
But
I
would
definitely
encourage
you
to
to
you
know,
follow
one
of
these
links
here
and
and
go
visit.
The
aisle
one
of
the
outland
pages
that
we
have
up
and-
and
you
know
you
can
peruse
those
pages
and
other
diagnosis
packages
that
we
have
on
the
website
that
allow
you
to
sort
of
get
a
first
sense
of
what
the
model
is
doing,
but
in
terms
of
the
islam
package,
there's
it.
You
know
assesses
these
a
big
range
of
variables
and
the
way
this
sort
of
works.
B
B
The
colors
indicate
whether
or
not
the
model
is
getting
better
against
all
these
different
metrics
that
are
involved
and.
C
B
Very
complicated
it's
hard
to
synthesize.
The
good
news
for
us
is
that
there
does
appear
to
be
some
sort
of
relatively
steady,
positive
trend
from
xenon
4
to
05,
but
it's
not
better
for
everything
in
some
variables
things
got
worse
and
I
could
definitely
talk
about
each
one
of
these
where
they
got
worse
and
discuss.
You
know
what
actually
you
know
what's
what
what,
where
the
source
of
that
of
the
degradation
or
improvement
has
come
from,
but
this
is
an
example
of
the
kind
of
tool
you
can
use
to
help
you
understand.
B
You
know
many
different
processes
are,
we
believe,
represented
better,
but
at
the
same
time
there's
many
more
moving
parts
and
additional
unconstrained
parameters
that
could
actually
degrade
to
grade
the
model,
and
so
there
is
nothing
saying
that
you
will
definitely
make
a
better
model
by
by
adding
new
improvements.
So
you
have
to
check
out
another
way.
You
know
another
question:
if
you're
trying
to
understand
whether
or
not
your
model
is
for
purpose
is
asking
questions
like.
Is
it
responding
to
environmental
perturbations
in
a
way
that
you
might
expect?
B
We
use
these
models
a
lot
to
study
climate
change
or
to
response
to
co2
fertilization,
but
it
isn't
while
doing
anything
sensible,
and
so
you
can
run
experiments
in
the
model
that
sort
of
replicate
real
world
experiments
like
the
sphere
of
carbon
exchange,
experiment
or
nutrient
addition,
experiments,
and
we
could
repeat
this
in
clm
and
add
new
nitrogen
or
add
co2,
and
then
check
out
the
response
and
I'm
gonna.
B
I
think,
skip
over
this
one
in
the
interest
of
time
and
look
at
this
figure,
which
shows
that
you
know
that
the
response,
so
you
can
look
at
the
on
the
on
the
y-axis
here-
is
the
co2
response
to
a
plus
200
ppm
of
co2
and
the
gray
bars
are
observational
estimates
of
that.
So,
ideally,
you
would
be
somewhere
within
the
gray
bars
and
then
on
the
on
the
x-axis
is
the
nutrient
response.
B
So
what
you
get,
if
you
add
50
kilograms
of
nitrogen
per
hectare
per
year,
and
so
there's
also
uncertainty
there,
and
so
we
saw
in
clm4
that
the
model
was
too
responsive
to
nitrogen
and
not
responsible
enough
to
co2
in
colon
4.5
will
look
a
little
bit
better
and
then
clm5.
B
You
know
we
appear
to
be
lining
up
within
within
the
observational
estimates.
Now
I
want
to
put
a
lot
of
caution
on
this.
If
you
look
at
the
plant
functional
type
level,
the
picture
is
much
more
murky,
and
it
also
could
just
be
that
we
lucked
out
right
that
we
happen
to
have
a
parameter
set.
That
lies
us
in
that
in
that
realistic
level,
and
we
can
definitely
have
sin
seen
since
that
we
can
have
parameter
settings
that
would
that
would
you
know,
not
match
the
observations.
B
Okay,
so,
what's
going
to
be
in
cto76,
I'm
not
going
to
try
to
present
that
you
know.
Our
hope
is
that
it's
going
to
be
this
slick,
new,
electric
bike,
that
you
know
everything
works
perfectly.
I
didn't
come
up
with
the
analogy
photo
of
what
it
actually
is
going
to
look
like,
which
probably
going
to
be
much
more.
You
know
a
little
bit.
Quirky
things
will
work
well
and
something
some
things
won't
work
well,
but
this
is
always
our
target
to
build
the
best
bicycle,
and
this
analogy
they
can.
B
They
can
do
the
most
things,
carry
the
most
things,
move
the
fastest
and
all
these
things,
but
I
will
say
that
you
know
how
are
we
deciding
what
to
do?
You
know
we
have
these
central
research
questions,
which
I
kind
of
highlighted
at
the
beginning,
that
kind
of
drive
our
development
priorities.
You
know
questions
about
ecosystem
vulnerability,
impacts
on
the
carbons,
likely
system,
services,
land-based
mitigation,
solutions,
water
and
food
security
and
sources
of
predictability
and
persistent
prediction.
B
You
know
one
of
the
big
topics
and
the
big
focal
points
has
been
over
the
last
several
years
is-
is
a
transition
to
what
we
call
faith.
The
functionally
assembled
trusted
ecosystem
simulator,
because
we
think
it
will
be
a
more
realistic
way
to
represent
ecosystem
functioning
and
how
ecosystems
are
going
to
respond
to
climate
change
as
well
as
natural
and
human
disturbances.
B
So
we're
really
excited
about
this
new
capability,
and
I
think
you
will
be
running
a
faith
simulation,
a
single
point
during
this
tutorial.
I
can't
hundred
something
remember
so:
apologies
to
the
organizers
and
there's
other
things
we're
working
on
to
try
to
adjust
these
research
priorities,
increase,
sub,
good
realism,
water
and
land
management.
B
Things
like
representing
the
plant
can
it
be
more
realistically
and
uses
somography
and
you'll
hear
a
few
about
a
few
of
these
things
and
and
during
the
lectures
tomorrow,
although
it
would
just
be
snippets
of
of
in
these
lectures
tomorrow,
because
you
know,
basically,
we
just
don't
have
the
time
to
go
through
everything,
but
I'll
get
to
the
end
about
ways
for
you
to
to
learn
more
and
to
be
to
become
more
deeply
involved
with
our
with
our
activities,
which
helps
you,
you
know
stay
on
top
of
all
the
latest
things
that
were
that
we
are
working
on
in
other
communities,
okay,
so
in
the
last
five
six
seven
minutes,
or
so
I
want
to
introduce
some
of
the
things
that
ctstem
can
do.
B
That.
I
think,
will
help
you
as
you,
you
know,
think
about
using
ctsm
in
your
research
in.
In
addition
to
the
you
know,
the
process,
complexity
and
the
subgrid
scale
complexity
representations
that
we
that
we've
included
in
the
model.
We
also
have
a
lot
of
flexibility
in
terms
of
how
the
model
is,
is
configured
and
run,
and
that
really
helps
you.
You
know
address
specific
scientific
questions.
So
we
have,
for
example,
many
different
model
configurations
which
I
list
here.
B
The
simplest
one
is
what
we
call
sp
for
satellite
phenology
and
this
in
this
mode.
We
prescribe
the
vegetation
state
from
from
observations,
modus
observations
and
so
in
this
mode
tends
to
be
the
most
realistic
in
terms
of
the
fluxes
you
get
on
the
model,
the
hydrology
you
get
out
of
the
model
because
you're
not
having
the
you
know
the
significant
degrees
of
freedom
from
from
a
prognostic
chemistry
version.
So.
C
B
C
B
Configuration
to
use
for,
if
you're
interested
in
hydrology,
if
you're
in
high
resolution,
we
tend
to
run
with
this
with
this
version,
because
it's
hard
to
spin
up
the
body
chemistry
version,
there's
a
biochemistry
version,
which
is
our
you
know,
calculates
prognostic,
carbon
and
nutrient
cycles
and
vegetation
state
is
prognostic
great.
For
studies
of
of
how
you
know
climate
interacts
with
with
the
carbon
cycle,
there's
a
bgc
crop
version,
which
is
the
default
version
we
have
in
csm2,
which
is
the
same
as
the
bgc.
B
But
with
a
with
a
prognostic
crop
model,
there
is
a
version
called.
We
call
bgc,
no
anthro
which
people
use
for
for
paleo
climate
configurations
where
basically,
we've
turned
off
all
of
the
anthropogenic
impacts
or
influences
on
on
the
system.
B
So
there's
no
there's
no
urban
areas,
there's
no
croplands,
there's
no
fire
suppression
or
ignition
from
humans
and
then
there's
the
fates
model,
the
phase
configuration
of
which
is
which
is
like
the
bgc
version,
but
replacing
the
vegetation
model
with
with
the
fades
vegetation
model
and
there's
also
many
options
for
individual
primarizations,
which
you
can
turn
on
and
turn
off,
and
you
really
just
have
to
get
into
the
model
to
understand
what
you
can
do
here.
B
You
can,
for
example,
revert
back
to
the
configuration
we
use
for
xeon
4.5
and
ask
questions
about
you
know
how
is
the
model
improve
over
time
or
maybe
there's
some
process
insane
4.5,
which
you
prefer
you
could
you
could
recover
that
process
and
turn
it
on
in
your
simulations?
B
We
can
obviously
run
global.
That's
where
our
target
is
low
and
also
high
resolution.
B
We
are
working
towards
the
capability
to
run,
or
at
least
have
run
short
periods
of
time
at
you
know,
global
high
resolution
at
three
kilometer
resolution
in
coupled
csm
simulations,
you
can
run
regional
simulations
also
at
high
resolution
or
low
resolution
single
point
simulations
at
tower
sites
or
any
site
around
around
the
globe
and
also
on
irregular
grids
like
the
cube
sphere,
which
tends
to
be
what
you
need
to
do
as
you
go
to
high
relations
or
even
a
catchment
grid,
which
we
have
shown
ourselves
to
be
able
to
do.
B
There's
also
different
modes
of
forcing
so
there's
the
we
call
the
anomaly
forcing
mechanism,
and
I
think
this
is
an
instant
one
and
people
are
using
it
more
and
more
and
basically,
in
this
mechanism,
you
can
use
just
monthly
anomalies
taken
from
a
from
an
ursa
model
simulation
like
csm
sort
of
add
them
to
cycled.
Reanalysis
data
like
we
normally
use
for
clm
to
look
at
the
impact
of
climate
change
on
land
processes
for
several
different
ssp
rcp
combinations,
and
we
have
four
ssps
that
are
available
essentially
out
of
the
box.
B
I
think
that
broke
in
the
latest
version
due
to
due
to
some
modifications
that
we're
working
on
to
fix
it,
but
essentially
it's
available
out
of
the
box,
and
so
this
allows
you
to
you
know
you
don't
have
to
run
the
full
csm
to
try
to
understand
the
climate
change
impacts
on
the
on
the
terrestrial
fluxes
of
water
or
carbon
or
energy.
You
can
just
use
this
anomaly
force
method.
B
We
have
several
different.
Forcing
data
sets
that
we
that
we
support
and
we've
seen
in
previous
research,
that
you
know
the
results
changed
depending
which
forcing
data
set.
You
use-
and
these
acronyms
may
or
not
mean
anything
to
you,
but
several
these
are
global.
We
also
have
the
regional
data
set
called
the
nldas
north
american
land
data
simulation
system.
B
We
have
the
ability
to
run
with
prescribed
soil
moisture
so
that
helps
you
unders.
You
know
to
do
research
on
questions
about
trying
to
understand
what
soil
moisture
impacts
are
on
climate.
It
often
helps
to
to
sort
of
be
able
to
control
that
soil
moisture.
We
have
alternate
land
use,
land
cover,
change,
data
sets
and
then
sort
of
realize
here
at
the
end
that
I
wasn't
even
mentioning
the
biggest
one
is
you
can
couple
the
model
to
cam,
csm
and
also
to
war
for
other
regional
models?
B
B
So
a
lot
of
different,
flexible
ways
to
to
sort
of
help
you,
you
know,
interact
with
the
model
to
address
the
questions
you're
trying
to
address,
and
so
then
I
want
to
finish
here
with
with
a
couple
of
things
that
I
think
are
important
to
think
about
as
you're
starting
off
on
your
c
teaches
summer
search,
I
think
gordon
bone
is
the
first
person
who
made
a
slide
like
this,
and
I
like
the
way
he
put
it.
He
said
cgsm
and
csm
is
just
a
starting
point
for
your
science.
B
It
is
not
the
science
itself,
it
is
really
easy
to
run
the
model
and
get
an
answer.
In
fact,
we're
going
to
show
you
today,
it's
it's
trivially,
easy,
almost
too
easy
to
run
the
model
and
get
an
answer.
It
is
much
much
harder
to
understand
why
you
got
that
answer.
B
The
model
is,
is
very
complex.
In
fact,
you
know
to
understand
why
why
I
did
something
is
it
is
a
huge
part
of
the
time
of
what
we
spend
here
and
in
fact
we
often
don't
understand
why
the
model
does
what
it
does
and
we
have.
You
know
to
work
with
each
other.
Colleagues
around
the
country
in
the
world
to
understand.
B
What's
happening
is
also
much
harder
to
improve
the
model.
If
the
model's
not
doing
what
you
like,
which
most
likely
out
of
the
box,
it
won't
be
operating
the
way
you
you
hope,
there'll
be
biases,
there'll,
be
limitations
and
trying
to
improve
the
model
is
definitely
highly
recommended.
That's
how
we
get
to
a
better
model,
but
it's
it's
hard
to
do,
and
I
like
this
slide
here
from
this
figure
from
from
a
paper
that
just
came
out
by
emily
kuiper
snowman
who's.
B
You
know
trying
to
interface
with
ecologists,
to
understand
how
ecologists
can
help
improve
global
resistance
models,
and
she
has
a
nice
way
of
putting
this
here.
You
know
the
illusion
is
that
you
have
data,
and
then
you,
you
know
you
just
do
some
modeling
and
then
you
have
some
insights
and
it's
just
easy
easy
as
pie.
The
reality
is
it's
a
really
complex
process
set
of
processes
and
interactions
to
go
from
an
idea
about
an
ecological,
behavior
or
hydrological
behavior
that
you've
observed
in
the
ear
system
and
transitioning
that
into
nurses
to
model.
B
I'm
not
going
to
try
to
go
through
this,
but
I
highly
recommend
reading
this
paper,
because
that
really
helps
you
kind
of
understand
how
how
you
work
to
build
a
model,
a
better
model,
and
then
the
final
point
I
want
to
make
here
is
that
it's
much
harder
to
design
while
experiments
to
help
you
learn
something
you
know
having
a
model
and
running
it.
We'll
show
you
how
to
do
that
today,
but
that's
really
just
the
first
step.
Then
you
have
to
think
about.
B
Let's
just
say
your
question
is
research
questions
about
how
land
use
change
has
impacted
climate,
so
a
relatively
straightforward
thing
to
do
is
to
run
two
simulations
one
with
and
one
without
land
use
change,
and
you
can
look
difference
between
those
simulations
and
get
a
sense
of
how
is
land
use,
change,
affecting
surface
fluxes
or
a
carbon
cycle.
B
So
that's
first
question,
but
then
you
might
say
well,
I
actually
want
to
learn
more.
I
want
to
learn
whether
you
know
whether
the
response
I'm
seeing
in
a
fully
coupled
model
is
due
to
the
atmosphere
or
due
to
the
land,
and
so
you
could
run
the
same
set
of
simulations
coupled
you
know
like
a
cmp6
historical
simulation
and
on
also
written
the
same
simulation
in
land
only
and
setting
these
two
simulations
together.
You
can
understand
you
know
what
is
the
foreseen?
What
is
the
response?
What
is
contributing
enforcement
response?
B
So
this
is
my
final
slide
and
say
you
know:
you're,
welcome
and
encouraged
to
get
involved
in
ctsm
activities.
I
think
it's
a
much
more
rich
experience
when
you
work
with
others,
so
we
have
many
ways
to
do
that.
We
have
the
csm
lab
model
working
group
workshops
both
in
the
winter
focus
on
the
landlord
in
isolation
and
also
the
summer
workshop
was
coming
up
in
a
few
weeks,
which
is
looked
at
the
fully
coupled
model.
B
We
have
a
github
open
community
development
platform
where
you
can
track
what
we're
doing
contribute
to
what
we're
doing
review
all
the
issues
that
we're
considering,
there's
bi-weekly
development
and
science
meetings
that
that
we
are
open
to
anyone,
who's
interested
in
participating
and
there's
opportunities
to
collaborate
with
other
scientists
that
are
working
with
chsm,
while
they're
ncar
or
we're
really
scientists
around
the
world.
And
so
we
really
encourage
you
to
to
get
involved,
and
that
is
my
final
slide.
A
A
If
you
do
have
a
question,
go
ahead
and
raise
your
hand
or
drop
it
in
the
chat,
as
dave
was
chat,
as
dave
was
presenting,
I
dropped.
You
know
a
link
to
some
islam
plots
linked
to
emily's
paper
and
the
sign
up
to
get
emails
on
the
lmwg
listserv.
If
you're
not
already
signed
up
for
those
notifications,
alan
go
ahead.
C
You
mentioned
fire:
does
the
model
actually
simulate
fires
and
change
the
vegetation
or
crops
and
emit
particles
and
and
gases,
and
interact
with
the
with
the
atmosphere
and
also
you
said
you
have
human
cause
fires,
but
does
the
model
react
to
thunderstorms
and
lightning.
B
Yeah,
it
does
react
to
thunderstorms
and
lightning,
although
currently
it's
a
prescribed
data
set-
and
I
think
we've
talked
about
this
before
we
were
talking
with
the
atmosphere
group-
about
getting
lightning
from
the
atmosphere
interactively
that,
hopefully
we'll
be
ready
for
the
next
version
of
the
model.
The
fire
emissions
to
the
atmosphere,
aerosols
and
co2
are
our
co2
does
get
immune
to
the
atmosphere.
B
When
we
run
an
emissions
driven
run,
aerosols
can
be
coupled
to
the
atmosphere,
it's
not
a
standard
configuration,
but
it
does
work
and
we're
studying
that
as
a
possibility
for
the
next
generation
model.
It's
a
complex
thing
I
can't
get
into
here.
Fire
does
it
does
affect
vegetation,
but
it
doesn't
affect
it
in
a
realistic
way.
Like
you
wouldn't
see
a
burn
scar
within
within
ctsm
within
fates,
you
would
so
that's.
B
One
of
the
advantages
of
moving
to
fates
is
that
it
would
actually
capture
the
albedo,
the
hydrology,
the
you
know,
the
other
processes
that
are
affected
by
fire,
which
are
essentially
not
captured
by
a
surprise,
so
obviously
I'd
be
given
an
entire.
Well,
probably
that
wouldn't
be
the
person's
one,
but
give
an
entire
lecture
on
the
fire
model.
A
Great
thanks
dave
there's
a
few
details
on
getting
folks
started
in
the
tutorial,
so
maybe
in
the
interest
of
time,
we'll
hold
off
on
further
questions
and
and
dave
and
others
will
be
around
to
answer
questions
over
the
next
couple
hours
as
we're
as
we're
starting
on
the
practical
session.
A
So
for
the
now
it's
like
total
switch
like
changing
gears,
I'm
going
to
share
my
screen
again,
but
if
you
all
could
get
on
to
the
cloud,
if
you
get
on
the
cesm
lab
on
the
cloud,
that
would
be
awesome
and
I'll.
Are
you
guys
seeing
my
screen
here?
A
A
A
A
And
hopefully,
when
you
log
in
you
can
get
to
a
screen
like
this.
So
if
you're,
not
if
you're
not
already
seeing
a
launcher,
you
can
click
on
this
little
plus
arrow
I'll.
Try
to
make
this
bigger.
So
it's
easier
to
see
so.
Click
on
the
plus
arrow
you'll
get
a
launcher
that
comes
up
and
then
navigate
down
to
open
a
terminal
window.
A
This
is,
if
you're
not
used
to
linux,
if
you're
not
used
to
jupiter,
notebooks
or
or
any
of
this
stuff,
it's
it's
going
to
be
kind
of
fast,
so
you
could
look
at
what
directory
you're
in
hopefully
you're
in
your
home
directory
and
cd
into
that
ctsm
tutorial
directory,
which
hopefully,
everyone
has
cloned
already
as
part
of
your
homework,
and
what
I'd
like
everyone
to
do
is
update
your
notebooks,
and
so
you
know,
since
many
of
you,
cloned
this
notebook
on
friday
or
saturday
or
monday,
or
yesterday,
we've
made
a
number
of
changes
to
the
day,
one
and
day
two
tutorials,
so
just
so
that
everyone's
working
from
a
local,
updated
local
copy.
A
A
And
you
should
see
something
on
your
screen
that
you
know
made
some
changes
to
a
number
of
files.
My
local
copy
is
already
up
to
date,
so
I
don't
have
anything
to
show
you
here,
but
that's
kind
of
the
first
order
thing
that
everyone
needs
to
do
is
to
to
have
an
updated
copy,
and
we
may
have
to
do
this
again
tomorrow,
depending
on.
If
we
find
find
more
challenges.
A
So
if
you're
having
troubles
doing
that,
maybe
stick
around
during
the
break
or
or
we'll
you
know
that
will
be
around
to
help
you
in
case.
That's
that's
you're,
having
problems
you're,
not
able
to
do
the
git
pull
so
in
trying
to
think
about
how
we
would
actually
make
this
tutorial
happen.
A
One
of
the
things
to
me,
one
of
the
things
that's
hardest
to
think
about
is,
is
you
know
what
normally
you
know
what
I'm
used
to
thinking
about
normally
happening
in
the
tutorial
and,
what's
actually
going
to
happen,
trying
to
do
this
virtually
on
zoom
and
so
normally
what
happens
for
the
practical
sessions?
A
You
know
we
all
go
to
the
library
or
we
all
go
to
the
damon
room
at
the
mesa
lab
kind
of
sit
around
big
tables
and
everyone
starts
in
on
the
practical
session,
and
when
people
have
questions
you
raise
your
hand
and
keith
or
danica
or
dave,
or
somebody
else
comes
over
to.
A
Pods
we're
gonna
put
everyone
into
breakout
rooms,
and
the
idea
is
that
you're
working
together
working
together
on
these
on
these
on
the
questions
and
on
the
on
the
you
know
the
cells
in
the
notebook
and
then
and
asking
questions
of
each
other
and
then
also
asking
other
people
for
help,
and
so
I
I've
kind
of
set
up
these
groups.
There
was
a
number
of
universities
that
kind
of
had
enough
people
together
that
it
kind
of
makes
sense
for
you
all
to
work
together.
A
So
learning
pod,
one
two
and
three
will
be
folks
from
rutgers,
yukon
and
cornell,
and
hopefully
that's
our.
Hopefully
your
labs,
if
rutgers
and
cornell
you're,
all
basically
in
the
same
lab
and
then
at
cornell.
It's
kind
of
interesting
there's
a
bunch
of
you
that
are
all
in
in
different
labs,
but
at
the
same
university
looking
at
different
questions
and
then
everyone
else
is
is
largely
lined
up
topically
and
so.
A
If
the
fates
team
all
wants
to
get
into
one
room
like
if
only
a
couple
of
couple
of
you
from
europe
are
gonna
are
gonna,
stay
on
and
you're
lonely
and
you
wanna
join
the
north
americans,
or
I
guess
the
north
americans
and
barbara
at
the
americas
on
the
main
fates
group.
That's
fine.
It
looked
like
a
number
of
folks
were
interested
in
data
simulation
and
with
machine
learning.
A
So
I
made
that
a
group,
some
others
on
ecohydrology
and
then
there
was
two
enough
for
kind
of
split
biogeochemistry
into
two
two
different
cohorts
and
so
by
geochemistry,
one
and
two,
and
then
we
also
set
up
a
study
hall
room,
which
is
for
those
of
you
who
haven't
had
a
chance
to
do
the
homework
yet,
which
I
think
is.
A
A
So
we've
got
about
a
10-minute
break
scheduled
and
at
10
after
the
hour.
What
I'd
like
to
ask
everyone
to
do
is
to
come
back,
join
your
breakout
room
and
and
then
go
ahead
and
introduce
yourself
to
the
group.
So
the
breakout
rooms
are,
you
know,
four
to
six
or
seven
people.
A
A
But
again
many
of
us
will
be
around
to
ask
for
help
or
for
you
to
ask
for
help,
and
if
you
do
have
questions
you
know
you
can
you'll
be
in
breakout
rooms,
but
you
can
send
somebody
out
into
the
main
room
to
ask
for
help
or
when
you're
in
the
breakout
room
you
can
click
on
the
breakout,
room,
icon
and
then
ask
for
help
and
that'll
ping,
elizabeth
and
others
that
are
in
the
main
room
about
who
you
know
asking
for
who
can
you
know
somebody
can
come
in
and
either
ask
help
with
a
technical
question,
help
with
a
scientific
question
or
anything
else
that
you
might
have
might
have
questions
for
so
please
make
use
of
the
resources
they're,
like
I
said,
there's
going
to
be
eight
to
ten
of
us
staff
that
are
around
and
and
yeah.
A
C
Yeah,
I
think
there
are
a
few
people
who
are
having
errors
with
just
that
get
pulled.
So
I
think.
C
If
we
could
have
maybe
a
breakout
room
just
for
folks
who
are
having
issues
with
that
first
step,
because
it's
seemingly
stemming
from
a
few
different
problems.
C
C
C
A
Cool:
let's
go
ahead
and
take
a
break.
I
can't
believe
we're
actually
on
time.
A
A
Yeah,
so
elizabeth
will
open
up
the
breakout
rooms
after
the
break
or
if
in
a
in
a
couple
of
minutes,
maybe
you
want
to
open
up
now.
How
do
we
want
to
do
this?
I'd
rather
like
I'd
like
to
let
people
have
a
break
before
they
feel
like
they
need
to
jump
in
the
rooms
and
yeah
when
you
when
we?
A
So
let's
do
that
when
we
come
back
from
break
in
10
minutes
you'll,
everyone
will
get
put
into
a
room,
and
so-
and
that's
that's
kind
of
been
done
by
behind
the
scenes
so
you'll
be
assigned
to
a
breakout
room
when
we
come
back
in
10
minutes.