►
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).
B
Creating
useful
and
usable
knowledge
in
turn
requires
that
we
are
connecting
with
people
that
need
and
will
actually
make
use
of.
The
information
that
we're
creating
and
saying
that
actionable
knowledge
is
useful
and
usable
might
not
be
very
enlightening
if
you're
not
familiar
with
the
literature,
so
actionable
knowledge.
B
We
can
also
define
actionable
scientists,
research,
that's
done
with
the
need
of
users,
end
users
under
consideration
or
in
mind,
and
usefulness
and
usability,
which
you'll
hear
me
use,
and
some
of
our
speakers
use
tend
to
be
the
cornerstones
of
actionable
knowledge,
where
usefulness
means
that
our
scientific
products
are
able
to
meet
certain
conditions
of
adequacy.
For
example,
the
knowledge
is
empirically
adequate.
B
The
models
demonstrate
representational
adequacy
with
respect
to
the
phenomena
of
interest,
and
the
information
being
produced
is
directly
relevant
to
be
able
to
supply
answers
to
the
problems
or
questions
that
the
end
users
might
have
and
quickly.
Note
that
this
understanding
of
useful
there's
some
ambiguity
in
the
literature,
as
it's
been
used
to
in
some
cases,
describe
the
gap
between
producers
and
users.
B
But
when
we're
talking
about
actionable
knowledge,
the
emphasis
is
not
only
necessarily
on
the
product.
The
process
by
which
we
come
to
know
and
produce
information
is
also
important.
There's
been
a
lot
of
talk
about
the
importance
of
communication
mediation,
especially
of
differences
and
approaches,
perspectives
and
values,
and
translation
across
boundaries
and
perspectives.
B
Equity
is
another
important
feature
of
the
process
of
knowledge
development,
and
these
can
actually
determine
the
assessments
of
the
usefulness
and
usability
of
knowledge.
For
example,
some
of
the
studies
that
have
been
done
demonstrate
that
engaging
in
communication
engaging
in
mediation
of
differences
and
directly
responding
to
them
and
translation
have
an
impact
on
the
credibility,
the
salience
and
the
perception
of
the
legitimacy
of
climate
information
and
agriculture
and
water
management
and
now
co-production,
which
is
probably
a
term
that
you've
heard,
is
one
process
in
the
literature.
B
B
B
Okay,
sorry,
my
slideshow
is
stopping
a
little
bit.
This
might
also
thinking
about
things
in
modeling
context
seem
like
a
bit
of
a
daunting
task
because
managing
the
development,
research
and
the
applications
of
a
model,
especially
a
state-of-the-art
and
complex
model
like
csm,
where
development,
research
and
analysis
are
distributed
across
many
different
disciplines,
groups
and
institutions.
It
might
be
like
well,
how
does
co-production
connecting
with
users,
especially
decision
makers
or
communities?
How
do
we?
How
do
we
go
about
doing
that?
B
And
I
think
we
actually
have
what
we
need
to
start
small
and
then
build
out.
There
are
several
researchers
in
our
community
that
have
made
or
are
making
connections
to,
find
out
more
about
users
needs
and
are
working
with
potential
user
communities
to
satisfy
some
of
the
criteria
of
usefulness
and
usability
or
at
least
investigating
those
criteria,
and
so
the
questions
that
I
previously
introduced
might
be
difficult
to
answer.
But
I
think
there's
some
immediate
questions
that
we
can
develop.
B
Answers
to
that
that
won't
feed
into
our
ability
to
answer
these
other
questions
about
actionability
and
a
gcm
or
esm
modeling
context,
one
of
which
was
actually
brought
up
by
gokan
in
tuesday's
discussion.
For
example.
Are
there
lessons
that
we
can
begin
to
learn
based
on
experience
of
members
of
the
community
to
take
a
turn
from
andy
newman
they're
super
users
who
are
already
actually
operating
at
the
boundary
between
producers
of
climate
data,
climate
data
or
climate
model
output
and
users?
B
How
can
we
engage
them
better
so
that
they
can
provide
us
with
insight
or
they
can
connect
us
up
with
these
user
communities
an
exam?
And,
lastly,
are
there
examples
that
we
can
readily
learn
from
to
begin
to
kind
of
map
out
an
approach
and
process
for
doing
this
type
of
work
and
to
jump
the
gun
a
little
bit?
The
answer
to
that
last
question
is
yes,
as
is
going
to
become
apparent
in
the
talks
we've
lined
up
for
the
session,
so
our
first
speaker
is
going
to
be
kevin.
B
B
B
And,
lastly,
before
I
hand
it
off
to
kevin,
I
like
to
remind
the
group
of
ncar's
code
of
conduct,
while
participating
in
the
session,
offer
constructive
feedback.
Consider
new
ideas
and
actively
listen
share
the
air,
especially
with
alternative
experiences,
perspectives
and
values,
show
appreciation,
acknowledge
teamwork
and
encourage
innovation.
C
All
right,
thank
you,
monica
and,
as
monica
mentioned,
I'm
going
to
talk
a
little
bit
about
some
work
that
we've
been
doing
to
use
cesm
as
a
tool
for
climate
change
event
attribution
before
I
get
started.
I
want
to
acknowledge
my
two
co-authors
for
this
specific
work
that
I'll
spend
most
of
my
time
on,
but
I
also
call
out
some
colleagues
throughout
the
presentation,
the
first
of
of
which
are
as
michael
weiner
at
lawrence
berkeley,
national
laboratory
as
well
as
colin
zarziki
at
penn
state.
C
So
the
main
motivation
for
this
kind
of
use
of
cesm
is
really
based
upon
the
fact
that
in
reality,
the
united
states
experiences
a
variety
of
different
weather
and
climate
disasters
every
year.
So
if
you
just
look
to
what
we've
all
probably
seen
to
some
extent
before
the
what
noah
puts
out,
which
is
the
billion
dollar
weather
disaster,
so
these
are
disasters
that
caused
over
a
billion
dollars
in
damage
just
for
2021.
C
Why
is
this
important?
It's
because
if
you
actually
look
over
kind
of
at
the
evolution
of
the
number
of
billion
dollar
disasters,
when
they've
been
inflation
adjusted,
you
can
see
that
there
has
been
a
a
stark
increase
in
the
number
of
billion
dollar
disasters
since
the
the
1980s
and
in
particular,
you'll
notice.
That,
of
course,
there's
still
a
diversity
in
where
those
disasters
come
from.
But
we're
going
to
focus.
C
This
talk
really
a
lot
on
the
yellow
bars
that
are
showing
up
here,
which
is
on
tropical
cyclones
and
so
you'll
notice.
The
last
few
years,
particularly
the
last
two
years,
have
been
particularly
damaging
from
the
number
of
events
that
that
have
occurred
due
to
hurricanes,
and
so
this
is
an
interesting
area
for
actionable
science,
because
it
suggests
that
there
might
be
a
connection
between
extreme
events
and
climate
change.
C
Now,
of
course,
if
you
look
at
a
lot
of
the
assessment
reports
that
come
out,
for
example,
the
national
climate
assessment
you'll
see
statements
like
this.
That's
shown
on
the
left
that
that
changes
in
extreme
weather
events
are
the
primary
way
in
which
most
people
will
experience
climate
change
and
that
human-induced
climate
change
has
already
changed
the
number
or
the
strength
of
some
of
these
events.
C
So
it's
really
just
the
middle
column,
divided
by
the
left
column,
and
so
what
we
typically
do
when
we're
using
a
climate
model.
We
run
then
future
projections
under
whether
it's
rcp
4.5,
which
is
the
middle
row
or
rcp
8.5,
which
is
the
bottom
row
in
which
we
run
these
same
simulations
under
future
conditions,
and
then
we
try
to
say
how
are
hurricanes
or
how
is
hurricane
rainfall
if
we
want
to
focus
on
the
hazard
the
that
leads
to
flooding,
you
know
how
are
those
changing
with
climate,
and
we
typically
produce
plots
like
this
right.
C
C
C
It's
worth
mentioning
that
in
this
case,
we're
outsourcing
our
data
assimilation,
so
we're
actually
using
the
the
initial
conditions
that
go
into
the
operational
global
forecast
system
from
noaa
and
we're
we're
running
these
at
various
times
for
specific,
real
storms
that
have
happened
and
we
call
these
our
actual
forecast.
So
this
is
our
our
best
attempt
to
see
how
the
model
does
at
simulating
not
only
the
intensity
but
the
track
and
rainfall
of
actual
observed
storms
similar
to
what
gfs
does.
C
But
then
we
also
use
the
fact
that
we're
working
within
a
climate
model
framework
to
actually
develop
a
counterfactual
case.
This
is
in
which
we
take
those
same
actual
forecasts
right.
So
this
is
sticking
with
the
storyline
approach,
so
we're
focusing
on
individual
events,
but
we
now
remove
the
large-scale
climate
change
signal
from
not
only
the
initial
conditions
of
of
those
forecasts,
but
also
from
the
boundary
conditions
or
the
sea
surface
temperature.
C
But
for
this
and
then
I'll
get
a
little
bit
more
detail
to
how
we
actually
calculate
these
these
counterfactuals
in
a
second.
But
I
want
to
highlight
one
of
our
first
examples
of
doing
this.
Oh,
this
is
just
a
quick
plug
for
the
fact
that
we're
using
the
beta
cast
implementation
that
has
been
developed
by
colin
sarzuki,
which
is
available
on
github,
and
I'm
happy
to
chat
more
about
that
in
offline
afterwards.
C
So
here's
one
of
the
first
cases
in
which
we
use
this
storyline
approach
within
csm.
It
was
for
hurricane
florence
in
2018,
and
this
is
work
that
was
published
in
science
advances
and
what
we
did
is
we
initialized
a
hundred
forecasts
at
911,
zero
z
in
advance
of
multiple
days
in
advance
of
hurricane
florence,
making
landfall,
and
then
we
initialize
those
actual
in
red
and
those
counter
factual
forecasts.
C
Now,
of
course,
after
landfall,
the
the
track
deviates
somewhat
from
the
simulated
tracks,
and
that
was
true
of
any
operational
model
actually
at
this
initialization
time.
But
it
does
suggest
that
the
model
is
fit
for
perfect,
because
it's
able
to
simulate
the
the
location
of
landfall
reasonably
well,
as
well
as
the
difference
in
the
tracks
between
the
contra,
factual
and
actual,
are
small,
which
means
we
can
do
an
apples-to-apples
comparison,
and
so
for
this
first
work.
C
It's
set
to
1850
conditions.
You
can
see
that
the
the
amount
of
rainfall
has
actually
really
decreased,
and
so
this
got
us
thinking
right.
So
if
we
can
do
this
for
one
storm,
can
we
automate
this
so
that
it
can
be
more
useful,
more
broadly,
and
so
what
we
did
is
we
used
it
for
the
2020
hurricane
season?
C
It
was
particularly
damaging
because
there
was
over
30
named
storms
in
the
north
atlantic,
many
of
which
made
landfall
in
the
united
states
and
were
billion
dollar
disasters
for
united
states
and
and
in
this
case,
we're
going
to
build
our
counter
factual
to
kind
of
mimic
somewhat
to
the
what's
done
in
the
ipcc,
in
which
we're
we're
focusing
on
what
happened.
The
observed
climate
and
then
we're
going
to
compare
it
to
what
would
have
happened
in
the
in
2020
if
it
wasn't
for
the
increases
in
in
human
greenhouse
gases
and
other
changes.
C
So
we
kind
of
focus
on
a
natural
forecast.
But
to
do
this,
we're
actually
using
the
sorry
the
projection.
But
to
do
this
we're
actually
using
the
csm
large
ensemble.
So
we
don't
have
the
the
all
his
first,
the
the
net
hist
in
this
case
we're
actually
going
to
compare
the
2020
conditions
to
the
1850
control
run.
And
so
we
use
that
to
calculate
a
perturbation,
which
is
so
it's
shown
here
for
the
sea
surface
temperature.
C
But
we
also
do
this
for
3d
temperature
and
specific
humidity,
and
we
apply
that
to
our
initial
conditions
and
we
rerun
forecasts.
So
this
was
our
first
attempt
to
do
this
operationally.
So
here's
just
I'm
going
to
walk
through
some
of
the
quick
results,
and-
and
so
this
is
now
a
forecast
initialized
every
three
days.
We
call
them
minecasts
because
we
did
it
after
the
season.
We
didn't
do
it
in
real
time,
but
you
could
do
this
in
real
time.
C
We
initialized
in
every
three
days
starting
on
the
first,
which
is
the
start
of
the
hurricane
season,
all
the
way
through
the
end
of
november,
and
then
we
we
keep
storms
that
form
in
the
model
that
are
within
some
set
distance
of
the
observed
storm.
So
we're
doing
a
little
a
selection
of
storms
to
make
sure
that
they're
well
simulated,
which
is
kind
of
similar.
C
What
I
had
mentioned
for
the
the
prototype
case
with
florence
and
then
we
produce
so
we're
producing
over
60
initialization
times
we
have
20
ensemble
members
each
time
and
so
we're
producing
over
1400
simulations
for
both
the
counterfactual
and
actual
and
then
we're
allowed
to.
Then
track
them
and
then
we
pull
out
the
tropical
cyclone
precipitation,
and
so
this
counterfactual
is
the
world
that
would
have
happened
or,
let's
say
the
2020
season.
C
That
would
have
happened
if
everything
was
the
same,
except
the
large-scale
climate
change
signal
was
of
what
it
would
have
been
in
1850
right.
So
it's
been
so
all
the
warming
to
date
has
been
removed
in
the
thermodynamic
signal,
and
then
of
course
you
look
here
and
you
say:
okay.
Well,
there
isn't
much
difference
between
those
precipitation.
You
know
how
could
this
actually
be
useful?
C
Well,
it's
because
it's
actually
the
extreme
amounts
associated
with
hurricane
seasons
that
have
the
particular
amount
of
damage,
and
so,
if
you
look
at
the
three
hourly
rain
intensity
and
all
of
the
ensembles
in
which
there's
hurricanes,
you
can
see
that
if
you
compare
the
red
line
to
the
blue
line,
that
the
red
line
has
always
shifted
above.
So
that
means
that
there's
always
a
higher
probability
of
these
extreme
three
hourly
rain
rates
that
are
approaching
two
inches
of
rain
in
three
hours
have
actually
increased.
C
And
if
you
look
at
the
99th
percentile
of
those
events,
we
actually
see
a
shift
in
that
distribution
throughout
the
entire
season.
No
matter
if
it's
a
strong
or
weak
storm,
and
we
see
a
shift
of
about
10
percent
so
about
10
increase
and
the
99th
percentile,
three
hourly
rainfall
rates
can
be
attributed
to
to
climate
change
and
if
you
actually
focus
not
just
on
all
storms
but
on
hurricanes,
you
actually
see
that
this
number
goes
up
a
little
bit
to
11.
C
So
that
means
that
the
climate
change
signal
is
actually
a
little
stronger
in
stronger
storms.
We
can
also
do
the
same
for
accumulated
rainfall,
so
this
is
a
three-day
accumulated
amounts
for
a
given
location.
And
again
you
can
see
that
these
are
shifted
towards
higher
values
and
if
you
focus
on
the
accumulated
99,
percentile
you'll
also
see
that
they're
shifted.
But
here
it's
about
five
percent.
C
So
I
just
wanna
I'll
just
finish
with
the
summary
right.
So
this
work
suggests
that,
by
focusing
on
storylines
that
we
can
see
that
there's
an
increase
in
precipitation
amounts
in
the
2020
season
due
to
hurricanes
or
sorry
due
to
climate
change,
and
that
these
can
be
done
objectively,
which
suggests
that
cesm
could
actually
be
used
to
create
these
kind
of
counterfactuals
almost
all
the
time
right.
So
you
don't
need
to
do
it
for
the
2020
season.
C
B
D
B
C
Useful
compliment
to
the
kind
of
work
you're
you're
undertaking
here,
yeah,
so
you're,
right,
paul
that
the
you
know
that's
very
similar
to
the
pseudoglobal
warming
approach.
The
only
kind
of
difference
here
is
that
we're
using
a
climate
model
which
of
course
has
a
variable
resolution
grid,
but
it's
also
the
same
climate
model.
That's
created
the
climate
same
signal,
so
we
aren't
right.
C
So
that's
one
of
the
advantages
but
in
in
but
in
terms
of
methodology,
there's
a
lot
of
similarities,
and
so
we
actually
have
done
the
inverse,
which
is
kind
of
what
you're
suggesting.
So
we
have
done
it
to
do
the
pseudo
global
warming-
and
you
give
me
you
know
the
trick
right
where
I
actually
have
the
next
slide,
which
so
we've
done
this.
D
B
B
G
C
C
You
know
if
you're,
if
you
went
to
that
one
figure
that
had
all
the
tracks-
and
you
count-
there's
actually
a
little
less
tracks
actually
in
the
present
climate
compared
to
the
the
1850,
the
counterfactual,
which
is
in
part,
be
coming
from.
In
some
of
these
cases
simulation
or
in
some
of
these
cases,
when
there's
a
storm,
that's
weekly
in
the
initial
condition,
it
won't
spin
up
in
the
actual
case
and
we'll
spin
up
more
actually
in
the
counter
factual
because
of
the
stratification.
Okay,.
A
A
C
That's
true
and
the
way
we
do
the
methodology
is
we
only
focus
on
three-day
chunks
so
that
there's
no
overlap
with
all
the
forecasts,
so
we
basically
create
a
re-analysis
essentially
right
of
the
full
season
and
it's
it's.
There
are
different
forms
of
when
they
spin
up,
but
some
of
them
do
spin
up
like
in
day
one
or
two
right.
They
weren't
there
in
the
initial
condition,
but
they
do
abruptly
spin
up.
B
G
G
Sorry
about
that
morgan,
I
will
apologize
profusely
if
you're
listening
either
way.
Anyway,
we
are,
let's
see,
can
I
move
my
slide?
Yes
along
right,
so
just
by
quick
introduction
of
eppery,
we
are
a
non-profit
organization,
doing
research
in
the
public
interest,
we're
a
membership
organization.
Most
of
our
members
are
utilities
and
grid
operators
and
quasi
governmental
organizations
like
public
service
commissions-
and
we
are
aiming
to
you
know,
understand
everything
all
the
challenges
facing
the
electrical
grid,
how
to
make
it
more
reliable
and
cleaner
and
better
in
every
way.
G
We
have
a
particular
interest
these
days
in
understanding
the
relationship
between
climate
change
and
climate
change
policy
and
grid
strategy,
and
we
are
working
on
that
project
in
a
bunch
of
areas,
including
understanding
what
kind
of
flow
carbon
electrical
generation
resources
will
be
on
the
grid
over
time.
The
interaction
between
the
electrical
grid
and
economy,
wide
efforts
to
reduce
carbon
emissions.
So,
for
example,
uptake
of
electric
vehicles
is
not
you
know.
G
G
So
we
want
to
make
sure
that
we
are
working
to
improve
our
understanding
of
the
risks
presented
by
weather
phenomenon
in
the
present
climate
to
the
present
electrical
system,
and
from
that
basis
we
want
to
go
forward
to
develop,
documented
and
vetted
means
of
estimating
the
change
in
risk
due
to
predictable
changes
in
climate.
So
what
are
predictable
changes
in
climate?
G
Well,
there's
a
lot
of
work
going
on
now
in
predicting
the
climate
on
a
decadal
time
scale,
starting
from
observed
initial
conditions
in
the
data
assimilation,
sort
of
mode
so
running
climate
models,
as
though
you
were
running
a
forecast,
a
weather
forecast
where
you
update
the
initial
conditions
and
then
generate
a
forecast
for
the
next
decade
or
so.
But
of
course,
the
dominant
knowable.
G
Okay,
so
this
is
a
a
kind
of
wordy
but
thorough,
summary
slide
for
the
whole
climate
ready
program
and
I'm
not
going
to
go
through
it
in
too
much
detail.
But
I
leave
it
here
for
those
who
might
come
across
the
slides
later,
who
can
take
their
time
and
look
through
it?
But
essentially
the
climate
ready
program
is
divided
into
three
work
streams,
which
somewhat
reflect
like
the
three
modes
of
the
ipcc
report,
so
work
stream.
One
is
physical,
climate
data
and
guidance.
G
We
need
to
understand
what
kind
of
information
we
need
to
assess
the
risk
on
the
climate
system,
so
changes
in
observable,
sensible
weather
phenomenon
that
we
will
need
to
know
about
in
order
to
assist,
assess
changes
in
the
climate
in
the
risks
to
the
electrical
system
and
then
work
stream.
Two
is
actually
assessing
the
vulnerability
of
the
climate
system.
To
of
the
sorry,
the
electrical
generation
system
to
climate
risks
and
then
work
stream.
Three
is
what
do
we
do
about
it?
How
do
we
help
the
electrical
system?
G
The
people
responsible
for
running
the
electrical
system,
assess
these
risks
and
assess
options
for
responding
to
them?
Okay.
So
now
I'm
going
to
talk
a
little
bit
more
about
what
we
are
looking
for
from
the
climate,
modeling
and
climate
analysis
community
right.
So
we
need
to
characterize
the
hazards
to
the
electrical
system
from
the
present
climate
and
from
climate
change.
So
how
do
we?
How
are
we
going
to
do
that?
G
Well,
the
most
obvious
thing
we
can
do
is
we
can
look
at
existing
sources
of
data
and
run
those
through
models,
one
of
the
things
that
every
does
is
generate.
Comprehensive
economic,
technological
models
of
the
electric
power
system
and
I'll
talk
a
little
bit
more
in
the
in
a
few
slides
later
about
the
requirements
of
those
models.
But
we
need
lots
of
data
about
the
variability
of
the
things
that
affect
the
operation
of
the
electrical
system.
So
most
obviously,
temperature
determines
the
amount
of
electricity.
G
G
That's
currently
available
in
re-analysis
data
sets
that
extend
that
now
extend
back
easily
50
years
and
more
is
that
we
have
about
50
years
of
forced
climate
change
under
our
belt.
So
in
thinking
about
what
might
happen
20
years
from
now
we're
30
years
from
now,
even
just
from
a
pure
data
perspective,
we're
not
operating
blind.
We
have
some
information
about
how
climate
change
is
already
working
itself
out
in
the
real
world
and
so
using
tech.
G
Okay,
so
in
terms
you
know,
once
we
have
a
sense
of
what
the
change
will
be
over
time,
then
we
can
start
looking
at
how
those
how
how
climate
variability
climate
change
and
climate
as
it
exists
now
impact
the
climate
system.
What
the
risks
are
to
the
things
that
we
care
about
coming
out
of
the
climate
system.
G
So
again,
this
is
a
kind
of
complicated
and
wordy
graph,
but
I'll
present
it
here
so
that
people
in
the
future
will
be
able
to
look
at
it
and
take
their
time
going
through
it.
But
what
we're
looking
at
here
are
what
are
the?
What
are
the
sensible
weather
impacts
on
various
components
of
the
system?
So
I
I
mentioned
a
couple
of
obvious
ones
before
when
it
gets
warmer
in
the
summer,
we
need
more
electricity
to
cool
our
homes
when
it
gets
warmer
in
the
winter.
G
We
need
less
electricity
to
heat
our
homes,
but
some
of
the
less
obvious
things
to
people
who
aren't
familiar
with
the
electrical
system
might
be
that
when
it
gets
warmer
thermal
generation
of
electricity,
so,
for
example,
nuclear
power
plants
that
generate
steam
and
then
that
steam
is
used
to
run
a
turbine
after
it
runs
a
turbine.
You
need
a
source
of
water
to
come
and
heat
up
and
run
through
the
turbines
again
and
often
the
way
that's
done
is
in
a
closed
system
where
we
evaporate.
G
We
use
evaporative
cooling
to
restore
a
a
working
fluid
back
to
a
liquid
state
and
run
it
through
the
power
system
again,
so
we
need
to.
We
need
to
be
able
to
cool
hot
stuff
in
order
to
generate
electricity,
and
that
cooling
is
obviously
less
effective
if
the
air
outside
is
warmer
and
more
humid.
Other
things
that
we're
concerned
about
would
be
impacts
of
extreme
events
on
the
electric
system,
so
ice
storms
that
might
knock
down
power
lines.
Will
those
get
more
frequent
or
less
frequent?
G
Okay,
so,
in
a
little
more
detail,
these
are
kind
of
the
top
line
variables
that
we
would
love
to
have
access
to
coming
out
of
out
of
climate
modeling
work,
and
I
I
put
in
red
here
or
no
dabene.
We
need
to
know,
especially
it's
it's
important.
We
want
to
communicate
that
to
do
this.
Well,
we
really
need
hourly
resolution
data.
Now
that
doesn't
mean
that
you
know
every
data
set
that
comes
out
of
a
gcm
needs
to
be
outputted
hourly
resolution.
That's
a
lot
of
data
and
we're
already
we
have.
G
You
know
we
have
data
management
problems
as
it
is,
but
we
need
at
least
enough
of
it
so
that
we
can
test
methodologies
to
go
from,
say
max
min
daily
temperature
to
realistic
time
varying
hourly
resolution
time,
series
of
temperature,
for
example,
and
there's
no
question
that
that
we
need
reasonable
amounts.
You
know
decades
or
at
a
time,
from
a
range
of
climate,
forcing
scenarios
of
hourly
data
8760
hours
a
year.
So
what
are
the
variables
of
most
interest?
G
Well,
as
I
mentioned
before,
temperature
humidity
at
wind,
speed
near
the
earth's
surface,
impact
electrical
demand.
They
like
they
impact
thermal
power
generation.
They
impact
photovoltaic
generation
because
solar
panels
are
more
efficient
when
they're,
cooler
and
they're
less
efficient
once
they
get
hot
and
temperature
also
impacts
transmission,
because
when
power
lines
get
hot,
they
tend
to
sag
and
if
they
sag
they
might
get
too
close
to
vegetation.
So
we
can't
let
them
sag
very
much.
G
So
we
have
to
back
off
on
the
amount
of
power
that
they're
transmitting
when
they
get
too
hot
downwelling
solar
radiation
and
its
direct
and
diffusive
components
will
impact
pv
electric
photovoltaic
electric
generation.
They'll
also
impact
electrical
demand.
If
you
have
bright
sun
shining
on
a
house,
it
takes
a
bit
more
air
conditioning
to
cool
it
off
wind
speeds,
not
only
at
the
surface
but
at
50
to
150
meters
impact
wind
generation.
G
We
need
to
know
the
wind
speed
aloft
to
do
a
good
job
of
predicting
how
wind
generation
will
fit
into
the
system
going
forward
looks
like
I
am
approaching
my
time.
So
I
will
accelerate
and
ice
and
snow
accumulation
impact
wind
and
solar
generation.
Icing
on
on
wind
turbine
blades
is
an
important
impact
in
parts
of
the
world
and
lightning
and
vegetation.
Dryness
relate
to
wildfires
that
impact
transmission
and
distribution
and
precipitation
and
evaporation,
changes,
impact
hydrological
generation
and
storage
potential,
as
well
as
biomass
production
potential.
So.
G
G
Okay,
briefly,
we
are
looking
at
risks
to
the
climate
system
that
the
things
that
electrical
generators
and
grid
operators
care
about,
especially
our
concepts
like
resource
adequacy
in
the
future.
Will
we
have
enough
electricity
to
meet
the
peak
demand,
and
so
I,
for
the
reasons
that
I've
discussed
above
climate
change,
is
going
to
impact
both
the
maximum
demand
that's
experienced
and
also
how
much
renewable
generation
will
be
happening
when
the
maximum
demand
is
experienced
and
similarly,
climate
change
may
affect
loss
of
load
episodes.
G
So
when
we,
you
know,
destroy
a
big
chunk
of
the
trans
emission
system
or
distribution
system
in
a
big
storm,
all
of
a
sudden,
the
load
goes
down,
but
electrical
systems
need
to
be
balanced.
The
amount
of
being
generated
at
any
one
time
has
to
be
very,
very
close
to
the
amount
being
used
or
you
can
cause
grave
damage
to
the
system.
G
G
Okay,
so
quickly,
this
is
again
for
for
future
consumers.
Of
the
of
this
talk,
you
can
read
through
this
in
detail,
but
electrical
system
planning
have
a
bunch
of
motivations
to
understand
this
stuff.
Those
motivations
are
gonna
change,
depending
on
where
you
sit.
If
you're
a
place,
if
you're
a
utility
that's
exposed
to
hurricanes,
that's
going
to
be
your
biggest
concern.
This
is
something
we
need
to
be
aware
of.
G
To
include
a
whole
bunch
of
stuff,
that's
not
directly
related
to
climate,
but
of
course
you
guys
are
familiar
with
that,
because
everyone,
everyone
understands
that
change
in
aerosol
composition
over
over
parts
of
the
world
that
currently
have
a
whole
lot
of
aerosol
pollution
is
expected
to.
So
the
aerosol
concentrations
are
expected
to
go
down
in
every
forcing
scenario,
because
we
understand
that
techn
that,
as
l
as
as
wealth
increases,
people
are
less
patient
with
having
horrible
air
quality
in
their
cities
than
they
wanted
to
go
down.
G
So
apart
from
climate
goals,
a
lot
of
things
are
going
to
change
that
are
going
to
impact,
how
our
models,
the
kinds
of
systems
we
should
expect
to
have,
and
that
will
be
affected
by
the
other
changes
that
we're
directly
talking
about
here
and
then.
Finally,
we
need
to
be
concerned
about
exactly
what
what
kind
of
planning
strategies
people
in
our
among
our
members
are
going
to
be
using
how
far
in
the
future
they're
willing
to
look
and
exactly
what
kind
of
risk
analysis
they're
they're
interested
in
performing
okay.
G
B
A
lot
great,
thank
you
dan
and
yeah.
Sorry
about
that
rush.
We
have
a
couple
of
other
speakers,
but
we
will
have
time
for
discussion
at
the
end.
So
if
you
have
any
questions
or
you'd
like
to
engage
with
dan,
we
can
do
that
later
and
then
also
your
powerpoint
will
be
made
available
online.
So
people
can
go
back
and
read
the
details,
which
is
great
because
there's
a
lot
of
very
important
information
on
those
slides
all
right.
B
B
E
I'm
peter
lawrence,
I'm
a
project
scientist
in
the
terrestrial
sciences
section
at
ncar
and
what
I'm
going
to
take
you
through
today
is
like
part
of
the
innovators
pro
program
and
a
project
that
I'm
involved
with
with
cleo
wolfie
hazard
and
he's
at
the
university
of
washington
and
he's
looking
at
looking
at
different
ways
of
approaching
climate
impacts
and
understanding
ecosystem
management,
especially
from
an
indigenous
perspective,
and
so
it's
sort
of
taking
the
models.
And
this
is
a
collaborative
effort
from
people
here
at
encar.
E
So
there's
andy
newman,
andy,
wood,
nike,
musakami
and
ethan
goodman
at
rao,
who
are
also
big
contributors
to
this
pro
project,
and
the
idea
was
basically
to
step
outside
of
the
normal
way.
We
do
science
and
pick
up
a
more
a
less
traditional
sort
of
scientist
role
and
becoming
more
involved
with
stakeholders
and
more
involved
with.
E
I
guess,
impacted
stakeholders.
That's
the
point!
So
we
have
this
collaboration
with
the
kuruk
tribe
and
ngos
in
the
in
in
the
klamath
river
basin
and
climate
river
basins
in
north
northern
california.
Southern
oregon
and
the
whole
idea
was
to
sort
of
look
at
indigenous
knowledge
and
protocols
and
and
try
to
understand
how
culturally
ecosystems
and
species
were
being
both
managed
and
utilized
and
protected.
E
E
So
there's
been
a
lot
of
benefits
in
doing
this.
We
we're
working
really
closely
with
these
stakeholders.
We
can
starting
to
test
our
models
at
sort
of
resolutions.
We
wouldn't
typically
use
them
at
we're,
also
working
closely
with
the
creek
tribe
now,
which
is,
it
was
fabulous
the
biggest
so
the
other
side
of
it
is
we've
got
some
serious
challenges
like
the
models.
Don't
really
have
current
past
or
future
representations
of
the
study
area.
E
They're
consistent
with
the
complexity
of
the
state
have
the
stakeholder
knowledge
and
we
have
15
pfts,
but
only
two
of
them
are
really
well
actually
only
two
trees
pfts
are
represented
in
this
region,
and
then
we
have.
We
have
an
incomplete
understanding
of
like
how
the
landscape
has
changed,
so
we're
looking
at
unique
ways
of
being
able
to
go
and
describe
land
use,
land
cover
change
and
fire
regimes.
E
So
from
our
perspective,
how
do
we
go
about
putting
together
land
use
and
land
cover
change,
but
we've
got
a
very
much
a
global
perspective
on
how
we
were
using
the
model
and
also
in
its
in
the
terms
of
deep
time.
So
this
is
sort
of
around
the
cmip
process,
and
so
what
we
do
is
we
take
a
lot
of
data
products
and
also
time
series
and
build
up
a
history
of
what
we
believe
the
land
surface
potentially
look
like
and
currently
looks
like
and
could
look
like
in
the
future.
E
So
you
can
see
here,
we
have
these
course
code
degree
grid
cells
that
we've
compiled
globally
and
trying
to
use
them
where,
on
the
left,
here
we
have
a
the
modis
land
cover
map.
You
can
see,
there's
a
sort
of
a
breakdown
between
the
two
of
them.
This
gets
even
worse
when
we
start
looking
at
irrigated
areas.
So
we
can
see
these
sort
of
artifacts
that
are
being
imprinted
onto
the
to
the
land
surface,
which
really
impact
the
hydrology
and
the
way
the
model
is
set
up.
E
So,
even
though
we
had
set
the
model
up
at
one
kilometer
resolution,
the
underlying
data
was
like
really
coarse
and
even
when
we
started
looking
at
some
other
data
sets
like
such
as
here
we've
got
the
natalie
evergreen
trees.
We
still
had
this
quite
a
degree
of
coarseness,
so
we
were
like
thinking
well.
How
are
we
going
to
address
this
problem
so
going
back
to
the
way
the
data
was
originally
developed?
E
A
lot
of
the
data
sets
that
go
into
this
process
are
actually
at
higher
resolutions,
one
kilometer
typically,
so
the
idea
was
well.
Can
we
take
what
we've
got
at
this
course
resolution
and
downscale
it
back
to
a
useful
resolution
and
the
short
story
is
yes,
we
can
do
that.
So
if
you
look
here
on
the
right
once
we
start
bringing
those
data
products
back
in,
we
can
keep
the
model
consistent
with
our
global
time
series
and
our
global
estimates
of
cropping
distributions.
E
And
when
we
start
looking
at
the
forests,
we
can
see
nearly
if
evergreen
trees
have
better
resolution
and
better
structure
as
well
and
finally,
looking
at
the
broadleaf
deciduous
trees.
These
are
areas
that
our
collaborators
and
stakeholders
thought
were
really
important,
because
these
are
areas
that
have
changed
dramatically.
E
E
E
We're
doing
this
comparison
with
the
the
original
data
that
I
was
showing
you
compared
to
this
new
high
resolution
data
and
the
next
steps
going
forward,
we're
starting
to
develop
we're
going
to
be
developing
these
land
use
histories
with
the
cr,
the
community
and
also
with
university
of
washington
collaborators,
so
being
able
to
take
historical
photos
and
being
able
to
take
place
names
and
be
able
to
talk
about
narratives
and
work
out.
How
would
that
narrative
can
be
turned
into
a
alternative
land,
land
distribution,
alternative
landscape?
E
How
does
a
model
work
at
this
resolution
and
we're
going
to
start
planning
how
we're
going
to
do
different
land
management
strategies
to
investigate
the
key
scientific
and
societal
questions
in
terms
of
climate
fire
and
ecosystem
management?
What
I
would
say
is
that
this
tool
is
actually
the
tools
that
we're
building
to
do.
B
B
D
E
That
would
be
great
yeah
and
yeah.
I've
seen
both
you
and
paulie
on
the
fate
talk,
the
fake
discussions
has
been
great,
so
yeah.
This
is
a
project
that
is
sort
of
like
I've
sort
of
been
brought
into
which
to
do
a
small
part
of
hopefully,
but
anything
that
you
can
you
can
provide
would
be
wonderful
like
because
this
is
this
is
where
we
start
challenging
the
models
at
resolutions
that
you
don't.
E
F
Yeah,
I
was
just
wondering
peter
I
you
know,
I
know
that
this
is
not
a
short
and
it's
also
a
lot
of
data
to
store.
You
know
if
we
thought
about
starting
to
run
or
like
store
data
at
these
high
super
high
resolutions,
so
I'm
just
kind
of
curious
how
what
it?
What
would
the
process
be
like
for
somebody
else
who
wanted
to
run
a
simulation
at
that
resolution
for
a
different
location?
Is
it
I
mean,
can
they
do
it
in
a
few
hours
or
do
they
does
it
need
you.
E
Right
so
so
I
guess
that's
the
we
want
to
general
well.
The
whole
point
was
to
build
a
generalizable
tool
that
could
be
used
anywhere
so
yeah
the
klamath.
The
klamath
was
a
really
great
sort
of
example
of
it,
because
we
have
a
lot
of
local
knowledge
and
a
lot
of
people
who
will
have
eyes
on
it,
which
is
great.
But
yes,
this
tool
is
is
very
generalizable
can
be
basically
set
up
because
it's
using
global
data
sets
to
do
that
down
scaling.
E
So
it's
using
the
higher
resolution
satellite
data
to
and
what's
what's
nice
is
that
you
can
see
both
against
the
google
earth
and
also
against
the
modus
data
sets
that
it
really
does
pick
up
that
sort
of
high
resolution
higher
resolution.
I
guess
one
kilometer
resolution,
as
opposed
to
these
sort,
of
course,
grid
cell
scales,
where
we're
running
the
model
with
typically
and
so
yeah.
That
was
that
was
actually
quite
a
a
relief
that
this
general,
a
general
sort
of
approach,
could
give
results
that
look
like
they're
quite
specific
to
a
region.
E
B
B
A
A
A
A
Cmip
is
at
the
foundation
of
us
having
historically
unique
knowledge
of
our
future,
which
gives
humanity
the
opportunity
to
affect
our
future
in
ways
to
establish
positive
outcomes.
I
will
make
the
case
here
that
cmip
simulations
are
not
the
right
type
of
simulations
for
adaptation
planning
to
frame.
My
argument
here
are
some
starting
points.
A
There
is
another
point
of
view
that
of
the
user
of
the
practitioner,
and
it
is
the
practitioner
who
defines
ultimately
usability
our
research
at
gleesa.
The
great
lakes,
integrated
sciences
and
assessment
center
investigates
the
gaps
between
scientists
notions
of
useful,
what
has
proved
to
be
usable
and
how
to
build
connections,
the
knowledge
system
between
knowledge
providers
and
users.
A
A
A
A
We
also
recognize
the
responsibility
of
scientists
to
describe
and
represent
uncertainty
in
a
way
that
neither
understates
nor
overstates
what
we
know
all
of
these
attributes.
Legitimacy,
credibility,
salience,
plausible
and
ethical-
are
to
build
trust
with
those
participating
in
a
particular
adaptation
problem
to
be
addressed.
A
A
good
lesson
is
from
weather
forecasting.
Global
weather
models
have
many
similarities,
with
global
climate
models
and
historically
were
construed
to
provide
deterministic
forecasts.
That
is
this
is
what
will
happen
increasingly.
Model
simulations
are
viewed
as
probabilistic,
not
deterministic,
and
provided
by
well-designed
ensembles
professional
weather.
Forecasters
do
not
view
the
model
as
providing
the
forecast
models,
provide
guidance
and
it's
to
be
used
in
combination
with
other
sources
of
information
to
provide
forecasts.
A
A
A
A
A
A
A
A
We
have
calculated
the
bias
of
65
models,
including
the
cement
models,
north
america,
cortex
models
and
the
university
of
wisconsin
models
that
are
our
primary
modeling
tool.
We
use
at
gleesa
shown
in
this
figure
are
biases
by
season
of
precipitation
from
minus
50
to
plus
50
and
temperature
from
minus
5
degrees
centigrade
to
plus
5
degrees
centigrade
white
is
unbiased
and
darker
colors
are
more
biased.
A
A
If
we
look
at
biases
and
temperature
at
five
degrees
and
a
hundred
percent
precipitation,
there
are
still
five
of
the
models
that
do
not
make
it
into
the
ensemble
bias
in
important
quantities
such
as
the
net
basin
supply
to
determine
water
levels
is
frequently
greater
than
100
percent
and
varies
month.
By
month.
A
Next,
I
want
to
look
at
this
issue
of
weather.
Above
I
mentioned
that
local
practitioners
are
attuned
to
weather,
they
know
the
summertime
flooding.
Events
in
the
great
lakes
come
and
organize
storms,
often
in
mesoscale
convective
complexes,
wind
storms
come
in
severe
thunderstorms,
direct
shows
and
tornadoes
global
climate
models
do
not
represent
these
events.
Regional
climate
models
cannot
credibly
generate
them
from
the
flawed
boundary
conditions
of
global
models.
A
A
A
A
It
is
perhaps
embarrassing
to
say.
Yes,
there
is
a
lake,
and
the
surface
temperature
of
the
lake
is
determined
by
interpolation
of
the
sea
surface
temperature
between
the
atlantic
and
pacific
oceans,
the
great
lakes,
more
generally,
local
details
must
be
present
in
any
credible
model
used
to
provide
guidance
for
adaptation
applications
in
the
great
lakes
regions,
the
overhead
of
determining
the
representation
of
the
north
american,
great
lakes
and
cement
models
is
far
too
large.
A
A
A
So
I
want
to
start
some
synthesis
and
conclusions
models
are
built
for
purpose.
Those
of
us
using
cement
simulations
in
adaptation
planning
are
using
models
that
were
not
designed
for
that
purpose.
Good
practice
of
the
science
of
modeling
tells
us
that
we
need
model
guidance
designed
for
the
purpose
of
adaptation.
A
Cement
models
were
built
for
understanding
by
comparison,
the
models
used
to
provide
guidance
for
weather
forecasting
have
quite
narrow
purpose.
They
are
designed
to
define
information
on
different
time
spans
at
different
spatial
scales.
They
are
focused
on
physical
processes
of
specific
phenomena
and
vulnerabilities.
A
A
A
B
Great
thank
you
richard,
and
I
think
he
pointed
out
some
really
interesting
features
of
models
that
we
need
to
pay
attention
to,
for
example,
that
models
are
built
for
purpose,
and
so
the
decisions
that
are
made
during
the
course
of
development
are
consistent
with
that
purpose
and
oftentimes.
Those
are
challenges
that
need
to
be
addressed
when
thinking
about
repurposing
them
for
things
that
the
decisions
are
not
consistent
with.
D
A
I
have
I'll
say
two
places
to
take
that
the
first
and
the
algorithms
that
we
use
to
work
with
practitioners,
we
consider
the
vulnerabilities
and
the
weather
associated
with
them
to
be
essentially
use
cases
and
then
minimally.
They
helped
to
define,
let's
say
the
the
compound
factors
that
a
community
will
actually
experience
and
therefore
they
provide
a
a
rational
foundation
on
which
to
look
at
inter
dependencies
and
the
development
of
scenarios
and
and
therefore
I
view
them
as
one
of
the
critical
elements
of
you
know
the
broader
knowledge
system
that
is
needed.
A
The
the
other
path
is
the
logic
at
gfdl
just
wrote
a
very
interesting
paper
on
complexity
and
models
and
the
role
of
machine
learning
and
the
role
of
analogs
looking
forward
and
made
what
I
think
is
really
a
fairly
compelling
argument
of
the
the
role
of
you
know
using
analogs
of
what's
happened
now
in
the
paper.
He
does
address
the
the
challenges
of
non-stationarity,
but
I
think
once
again
they
become
one,
a
very
important
aspect
of
looking
at,
inter
and
activities
and
two
when
you
consider
that
the
climate
is
non-stationary.
A
B
D
F
F
F
Well
then,
a
bit
surprising
at
that,
but
I'm
glad
that
it
worked
okay.
So
today
I'm
gonna
talk
to
you
a
little
bit
about
farm
management
decisions
and
the
cesm
crop
model,
and
let's
see
okay.
So
this
is
just
a
an
aerial
photo
of
agriculture,
and
what
I
want
to
illustrate
with
this
particular
photo
is
that
agriculture
changes
the
land
surface.
It
changes
how
it
looks
which
is
important
for
the
physical
impacts,
and
it
also
changes
how
it
functions,
which
is
important
for
the
biogeochemical
impacts.
F
Figure
that
illustrates
one
of
the
major
challenges
here
and
that's
the
mismatch
in
scales
on
the
left.
I
I
just
have
this
picture
of
it's.
It's
a
satellite
photo
really,
but
this
is
kind
of
the
an
approximation
of
the
scales
that
air
system
models
typically
represent.
There
are
larger
spatial
scales
at
relatively
coarse
grid
resolutions
like
peter
mentioned
in
his
earlier
talk,
and
these
you
know
at
this
earth
system
modeling
scale,
it's
useful
for
generalizing
processes,
thinking
about
regional
to
global
scale
and
understanding
climate
interactions.
F
Also
any
model
development
that
happens
must
be
globally
scalable.
So
that's
one
purpose
of
the
earth
system
models,
but
the
scale
of
interest
for
farmers
is
more
at
the
field
scale
and
currently
earth
system
models
are
not
very
good
at
down
scaling
to
this
field
scale
and
even
if
they
were,
it,
wouldn't
necessarily
be
accurate,
because
downscaled
to
meteorological
data
might
still
not
be
high
enough
resolution
to
capture
some
of
that
variability
on
the
field
scale.
We
also
might
not
have
the
right
input
data,
but
we
are
making
progress.
F
As
peter
pointed
out,
there
are
data
available
at
smaller
scales,
so
we
can
start
to
think
about
those
things
within
the
community
er
system
model.
We
represent
major
commercial
and
bioenergy
crops,
and
but
I
want
to
highlight
that
many
of
the
vegetables
and
tree
crops
that
smaller
scale
farmers
use
are
not
necessarily
represented
in
a
model
like
cesm.
F
Representation
of
crops
for
an
earth
system
model
is
still
going
to
be
very
limited
in
usability
for
some
of
these
smaller
scale
farmers,
just
because
we're
limited
in
the
types
of
crops
that
we
have
or
are
able
to
represent
at
this
point,
and
also
the
phenology
of
the
crops,
meaning
the
timing
of
those
changes
in
plant
growth
stages,
including
things
like
planting
and
harvesting.
Since
those
are
all
temperature
triggered
we're
not
accounting
for
some
of
these
other
decisions
that
farmers
are
making.
F
We
also
include
basic
representation
management
in
the
crop
model
in
cesm,
and
so
there
are
two
types
of
management
fertilization
is
prescribed
by
crop
type
year
and
by
country
and
irrigation
is
applied
daily
when
soil
moisture
is
below
a
fixed
threshold
value,
and
so
these
these
are
both.
These
are
two
of
the
biggest
management
processes
that
farmers
undertake,
but
they're
still
relatively
rudimentary
in
representation.
F
Like
I
said,
for
an
earth
system
model,
this
is
a
fantastic
advance,
but
when
you're
thinking
about
a
farmer
on
a
field
scale
trying
to
make
decisions,
it's
it's
quite
limited,
and
so
we
need
to
keep
thinking
about
what
other
types
of
management
are
farmers,
thinking
about
and
or
using.
So
I'm
going
to
highlight
just
very
quickly
one
project
in
the
innovators
program,
but
by
kelsey
emerge.
F
There
are
actually
two
innovators
projects
that
are
focusing
on
agriculture
and
the
innovators
program,
I
should
say,
is
a
program
to
connect
social
scientists
with
end
car
scientists
and
so
kelsey's.
Project
kelsey
is
from
oregon
state
university
and
her
project
objective
is
to
really
develop
transferable
knowledge
about
how
scientific
data
contributes
to
former
soil
management
decision
making
and
also
to
identify
factors
that
shape
farmer
use
of
climate
and
soils
data.
F
F
You
know
a
scale
mismatch
that
we
have
in
the
models,
so
you
can
see
the
resolution
of
the
model
grid
cell
and
then
the
red
dot
is
the
willamette
valley,
which
is
what
kelsey
is
focusing
on,
so
that
falls
within
a
single
grid
cell
and
even
a
single
grid
cell
within
cesamos
is
much
larger
than
the
willamette
valley,
and
so
this
is
one
example
of
data.
That's
being
used.
F
That
kelsey
is
using
in
farmer
interviews
and
it's
just
showing
soil,
moisture
july,
soil
moisture
in
the
1850s,
the
2010s
and
then
the
20,
90s
and
I'll
highlight
this
swim.
Moisture
data
on
the
next
slide
as
well.
So
the
approach
that
we're
using
is
to
use
cesm
lens
2
data,
so
that's
the
large
ensemble
data
from
cesm2
and
you
can
see
a
different
example
of
this
data.
Again.
F
This
is
soil
moisture
in
the
top
10
centimeters,
and
this
is
a
time
series
plot,
and
what
we
really
like
about
this
plot
is
that
is
that
it
can
show
uncertainty
as
well,
which
I
think
is
really
important
thing
to
highlight
farmers
and
so
with
the
lens
2
data.
We're
showing
time
series
like
this
we're
also
showing
maps
like.
I
just
showed
you
on
the
previous
slide
and
annual
cycle
for
relevant
climate
variables
and,
like
I
said
this,
you
know
time
series
like
this:
that's
using
the
large
ensemble
can
help
to
communicate.
F
F
She
has
done
this
for
a
different
project
using
python's
bokeh
tool,
which
allows
users
to
have
sort
of
a
an
interface
where
they
can
select
different
sites
different
variables,
different
time
scales,
those
kinds
of
things-
it's
actually
much
harder
to
do
using
the
lens
2
data,
because
there's
a
lot
more
data,
and
so
we're
still
in
the
process
of
trying
to
figure
that
out.
But
I
do
you
know.
I
just
want
to
highlight
this.
F
This
is
the
lens
2
data
are
data
that
are
available
to
the
scientific
community
and
we
can
make
them
more
usable,
and
so
I
want
to
you
know
just
challenge
the
cesm
community
to
think
about
how
to
make
those
more
usable
and
kelsey's
project
is
one
of
three
out
of
nine
funded
projects.
Nine
fully
funded
projects.
F
I
should
say
that
are
actually
using
lens
data
in
this
capacity,
and
so
these
kinds
of
data
are
really
useful
to
the
decision
making
community
and
we
need
to
work
with
them
more
to
find
out
what
variables,
what
time
scales
and
what
values
are
used
for
the
most
usable,
so
that
we
can
hopefully
provide
some
of
those
data
in
ways
that
people
can
can
explore
themselves
and
and
interact.
And
then
the
last
step
of
this
project
is
that
kelsey
is
going
to
run.
F
We
can
also
gain
insight
into
the
usability
of
cesm
products
to
inform
regional
and
local
community
practices,
and
that's
what
I
was
saying
with
the
lens
2
data
that
I
think
could
be
really
usable
in
that
capacity
and
last
there's
benefits
in
that
it
provides
better
knowledge
of
agricultural
community
needs,
and
so
that
can
help
us
inform
the
model
development
that
we're
doing
here
at
ncar.
There
are,
of
course,
with
any
of
these
projects.
There
are
multiple
challenges
I
highlighted
a
few
of
these.
F
F
You
know
that's
in
general,
but
also
using
a
model
like
cesm,
which
is
somewhat
limited
in
spatial
scale,
and
some
of
the
verb
rules
are
limited
to
portal
scale,
although
30
minute
time
resolution
for
the
model,
if
we
can
save
data
on
that
time
scale,
I
think
makes
these
reusable
so
I'll.
Stop
there
and
answer
any
questions.
B
F
B
F
You
know
mattv
is
here,
so
I
can
just
really
quickly
highlight
that
you
know
it.
I
guess
the
question
is:
wouldn't
it
be
interesting
not
just
to
ask
during
the
interviews
what
data
is
usable
but
also
about
what
temperature,
precipitation
or
soil
moisture
data
patterns
farmers
use
for
their
decision
making
to
improve
the
crop
model?
And
I
do
think
that
that
is
interesting
and
necessary.
F
F
B
D
D
B
D
B
B
With
the
tuesday
session,
this
is
a
cross-working
group
session,
but
there
has
been
some
interest
in
the
possibility
of
developing
a
working
group.
That's
dedicated
to
addressing
some
of
the
questions
that
were
brought
up
in
today's
session
and
also
some
of
the
questions
that
were
touched
upon,
touched
upon
and
are
related
that
were
discussed
in
tuesday's
session.
So
if
you
are
interested
in
seeing
that
happen,
and
you
would
like
to
participate
just
send
me
a
quick
email,
noting
your
interests
and
once
we
get
some
input,
we
will.
B
B
So
yeah,
my
email
is
there
in
the
middle
of
two
questions:
sorry
about
that,
it's
monica
mo
ucar.edu
and
julio.
We
can
talk
about
this
later,
but
as
richard,
and
I
think
somebody
else
mary
put
in
the
chat.
No,
no,
no,
no
high
resolution
might
be
a
feature
of
usefulness
and
usability,
but
there
are
a
plurality
of
factors
that
go
into
determining
what
by
me,
might
be
useful
and
usable
in
any
given
context
and
they
can
differ
across
the
context
too.
B
B
Yeah,
thank
you,
and
I
mean
danica
put
in
the
chat.
The
resolution
is
definitely
one
factor
in
considering
usefulness
and
usability,
as
are
like
investigating
what
user
spatiotemporal
scales
are.
So
these
are.
These
are
things
that
that,
when
you're
working
with
users
need
to
be,
they
need
to
be
directly
consulted
to
kind
of
figure
out.
B
A
You
know
that
was
in
your
backyard
and
even
had
a
state
climatologist
tell
me,
you
know,
produce
me
a
number
in
my
backyard.
It
doesn't
matter
whether
it's
accurate
or
not.
I
need
that
number
and-
and
so
you
know,
steve's
idea
was,
you
know,
point
was
you
know
there
really
are
perhaps
ethical
limits
to
to
what
we
can
do,
and
you
know
the
pursuit
you
know
of
what
bellagi
called
in
his
paper
charney's
ladder
and
descending
charney's
ladder
that
that
is
probably
not
especially
well
founded
so
but.
B
D
Yeah,
I
guess
I've
enjoyed
these
this
discussion,
the
discussion
and
presentations
earlier
in
the
week-
and
I
I
guess
I
was
wondering-
is
there
sort
of
an
interest
in
kind
of
reactivating
what
a
number
of
years
ago
was
the
kind
of
a
societal
mentions,
type
working
group.
You
know
where.
So
I
guess
I'm
just
sort
of
wondering
where
this
is
headed,
and
maybe
this
is
jumping
the
gun
a
little
bit.
D
You
know
some
of
the
challenges,
but
also
opportunities
for
the
folks
who
want
to
make
use
of
products
that
we
generate.
You
know
as
part
of
our
science
and
so,
but
I
also
know
that
that
program
went
away
for
a
reason
or
for
reasons
perhaps-
and
I
guess
I
was
just
wondering-
yeah,
where,
where
we
think
this
is
headed,
if
it
and
I,
if
that's
a
difficult
question
to
answer-
that's
fine,
we
don't
you
know
have
to
then.
B
So
I
in
I'm,
going
to
try
and
keep
the
momentum
going
and
we're
going
to
have
some
meetings
with
people
that
that
say
that
they're
interested,
but
I
think
it
it
will
to
some
extent,
be
left
up
to
the
community
to
see
whether
we
can
get
something
sustained
going
on
and
also
what
that
might
look
like,
because
these
are.
These
are
themes.
B
That
kind
of
I
presented
things
that
I
think
are
important
and
then
our
speakers
think
are
important,
but
there
might
be
additional
things
that
the
group
would
need
to
cover
so
kind
of
what
that
theme
might
be
might
emerge
out
of
future
conversations
yeah
and
richard.
I
think
you
had
your
hand
up,
but
you
put
it
down,
but
do
you
would
you
like
to
it.
B
B
B
No
okay!
Well
I
thank
you
again
for
sticking
around
to
converse
with
colleagues
about
these
matters
and
again
send
me
an
email.
We
do
actually
also
now
have
a
actionable
science
and
esm
session
accepted
at
agu,
so
that
advertisement
will
be
sent
out
to
the
csm
community
shortly,
and
so
I
know
that's
like
months
away.