►
From YouTube: SimPEG meeting June 25
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
How
to
follow
that
up?
Gosh
I
didn't
work
west,
so
I
should
quick
updates
and
then
anybody
else
has
quick
updates,
feel
free
to
jump
in.
We
merge
this
morning
a
pull
request
from
dieter,
which
updates
the
versions
codeinsight
Megan.
So
what
that
does
is,
if
you
put
it
in
the
end
of
a
notebook
or
the
end
of
a
script,
it'll
print.
A
What
Birdman's
of
like
base
software
that
you
also
the
version
of
stimpack
so,
for
example,
if
you're
publishing
research
note
something
like
that,
and
you
want
it
to
be
clearly
something
to
you
start
around.
You
really
lots
of
background
noise
thanks
yeah,
so
that
prints,
the
versions
of
like
matplotlib,
numpy,
sie,
pi,
pi,
matte,
silver
and
discretize,
as
well
as
simple.
A
So
if
you're,
publishing
research,
notebooks
and
like
just
want
people
to
know
what
software
you're
using,
you
can
add
that
at
the
end,
so
it
used
to
be
you
would
import
the
lowercase
versions,
and
now
it's
uppercase,
because
it's
actually
a
class
that
you're
calling
so
have
a
look
at
the
pull
request.
Notes
if
you
have
any
questions,
feel
free
to
ping
on
slack
and
dieter.
I
can
jump
in
and
help.
A
C
D
D
E
A
Is
there
a
way
that
you
guys
can
bring
those
changes
in
like
in
a
somewhat
small
pull
request,
I
realize
that
you've
been
working
off
of
some
branch,
but
are
there
like
bite-sized
things
that
you
can
bring
in
so
we
can
do
that
quickly.
If
it's
on
a
giant
branch,
it's
gonna
just
happen
floor.
That's
sort
of
your
things,
yeah.
B
B
D
D
Sorry
I
think
so
we
we
inverted
that
yes
IP
data
a
long
time
ago
and
then
recently
got
a
drilling
drill,
holes
so
yeah
we
were
comparing
and
then
like
a
first
order.
It
seems
like
matching
pretty
well
with
the
the
gold
si,
but
we're
not
100%
sure
yet
so
yeah
we're
in
the
process
of
confirming
it
so
yeah.
Well,
we'll
update
you
guys
later
yeah,
we're
not
sure
at
the
moment.
Yeah.
B
Okay,
thank
you,
so
I
just
wanted
to
share
a
bit
with
you.
What
are
the
first
results
of
the
research
here
and
so
get
some
feedback
on
what
I
did
to
its
impact,
so
maybe
to
get
you
started,
but
I
wanted
to
show
this,
because
that's
actually
I
really
need
to
not
call
it
quality,
but
this
is
what
we
are
looking
at.
Actually,
in
my
research,
it's
an
irrigated
fields,
potatoes
and
you
see
the
sprinkler
going
there
so
like
the
main
questions
that
are
present
here
are
about
how
much
water
should
be
used
is.
B
Are
we
arrogating
in
an
efficient
way
in
this
kind
of
systems?
And
if
not,
how
can
you
change
that?
Not
only
for
the
amount
of
water,
but
also
in
terms
of
leaching
of
solutes,
like
nitrates
and
in
Belgium
there's
we
are
like
they
call.
It's
a
protects,
an
area
where
we
have
an
enlarged
or
increased
risk
of
magnet
leaching,
and
so
farmers
have
to
be
really
careful
about
yeah,
putting
too
much
water
and
too
much
nutrients
on
their
fields.
So
this
is
the
context,
and
you
can
already
see
that
such
an
irrigation
is
quite
heterogeneous.
B
Also,
if
you
see
a
huge
spread
of
the
the
beam
actually
due
to
wind
and
also
the
fact
that
I
just
displays
this
this
thing
in
the
field
until
it's
entirely
covered,
so
we're
talking
about
negation
which
takes
like
2
days
to
do
one
immigration
of
the
fields,
so
it's
quite
labor
intensive
and
and
it's
because
farmers
in
Belgium
don't
have
to
irrigate
in
a
systematic
way.
They
do
it
when
it
becomes
to
try
and
have
some
kind
of
a
back-up
plan.
So
so
that's
the
question.
B
That's
any
of
the
the
data
that
I've
been
analyzing
here
is
how
can
we
relate
potato
fields
in
an
efficient
way,
or
is
it
happening
efficient
at
this
moment?
And
and
how
can
we
quantify
this?
Because
that's
the
big
question
you
can
see
already
you
have
this
typical
system
with
ridges
and
furrows
I,
don't
know
whether
it's
cultivated
in
the
same
way.
A
B
Which
makes
it
already
quite
complex
from
Hyderabad
or
a
hydrological
perspective
here:
where
does
the
water
go
that
you
actually
apply
on
the
system?
There's
slopes?
There
is
accumulation,
there's
the
canopy
that
gets
some
of
the
water
which
can
go
to
the
to
the
stem,
then
so
yeah.
How
can
we
quantify
this,
and
in
addition
to
that
here,
you
can
see
our
that
looks
from
the
insights.
B
So
there
is,
of
course,
already
some
quite
old
research,
even
though
in
this
in
this
respect,
you
can
imagine
because
it's
an
old
technique,
so
people
have
used
different
techniques
to
look
at
these
things
like,
for
example,
dye
tracing
I,
don't
know
whether
you've
already
seen
slices
like
that
they
use
these
food
stain
colors
like
building
blue
or
others
too,
and
they
apply
them
with
irrigation
water.
And
then
they
look
at
what's
the
pattern
of
this
blue
colour
in
the
soil
afterwards
by
digging
a
big
trench?
B
Basically
it's
very
simple
and
there
or
you
can
put
sensors
at
different
locations
in
a
trench
and
and
see
what
happens
after
irrigation.
That's
not
a
study
with
TDR,
but
the
problem
is
in
the
literature.
All
these
studies
give
very
different
results.
Some
say:
there's
a
infiltration
mainly
happens
in
these
four
hours,
and
so
a
lot
of
it
is
lost
because
it
directly
goes
to
the
grandmother.
Other
half
others
have
shown,
for
example,
with
this
diet
tracing
that
there
is
intercept
interception
of
this
water
and
it
is
efficient.
B
So
that's
quite
quite
a
controversy
and
not
three,
a
lot
of
yeah
quantitative
information
on
and
what's
really
up
in
there,
and
why
is
that?
Well,
that's
what
I
explained
is
actually
a
quiet,
difficult,
ideological
problem,
it's
mainly
the
boundary
condition
which
is
not
clear
here.
So
if
you
look
at
the
hydrological
simulation,
it's
a
bit
like
an
electrical
simulation
and
we
have
our
sis
equation
Richards
equation
and
we
have
to
set
the
table
boundary
condition
which
is
the
rainfall.
B
But,
as
we
know
in
this
system,
it's
not
like
there's
one
blue
line
which
is
continuous.
We
can
expect
that
there's
a
different
amount
of
water
that
is
arriving
on
the
top
and
on
the
sides
and
in
the
middle,
and
you
can
even
see
it's
more
complex.
If
you
look
at
the
bottom
corner
there,
there
is
most
probably
and
that's
already
quantified,
actually
the
leaves
that
are
capturing
a
lot
of
water
and
directly
to
the
stem
of
the
plant
and
they're
there.
B
That
could
be
punctually,
a
very
big
amount
of
water
arriving
at
some
irrigation
intensity.
So
that's
also
a
difference.
If
it's
and
just
a
small
rain,
the
distribution
will
be
different
than
when
you
we
have
a
big
jet
of
water
arriving
in
a
short
time,
so
that
that's
a
bit
the
problem
and
I
think
that
does
basically
our
idea
I
think
he
Artie
could
be
a
quite
nice
technique
to
show
in
2d
what's
happening
there
like
you,
can
easily
put
electrodes
on
the
surface.
B
Underwater
arrives
after
an
irrigation
event:
where
is
the
connectivity
increasing?
Because
that's
mainly,
you
were
later
than
to
the
increase
in
soil
moisture
in
this
case
and
yeah?
So
that's
why
what
we
did
so,
let's
hope,
have
an
ERT.
We
can
find,
or
at
least
make
a
qualitative
picture
of
where
things
are
moving
and
to
which
extent,
and
maybe
also
quantitatively,
because
we
also
have
better
physics,
of
course,
and
we
could
then
convert
it
into
real,
so
much,
that's
even
better,
not
only
just
an
image
of
timidly
changes
but
really
go
to
so
much.
B
But
and
I'll
come
to
that
later,
there's
a
lot
of,
let's
say,
complications
in
this
because
often
like
farmers
use
ground
water
for
irrigation,
so
it's
quite
cold
as
compared
to
soil
in
summer.
So
you
have
like
conflicting
changes
of
impacts
on
resistivity,
which
are
less
simple
than
it
may
seem,
but
in
general,
in
an
agricultural
context,
when
water
is
added
to
the
soil,
there's
immediately
a
lot
of
solutes
which
are
dissolved
or
transported,
and
so
water
is
generally
more
conductive
or
adding
connectivity
to
the
meeting.
F
B
Is
actually
the
data
you'll
see
at
the
end?
So
this
is
what
happens
after
irrigation
actually,
but
so
what's
the
point
before
doing
that,
so
that
it
has
been
done
already
like
you
might
hope,
you
did
it
on
several
cases
in
eco,
hydrological
context,
in
forests,
in
mixed
farming,
but
maybe
on
flat
surfaces
or
on
SCO's,
but
with
without
an
important
micro
topography
like
this,
and
so
how
reliable
will
our
measure
measurement
be?
In
this
context?
That's
actually
the
question
that
I've
been
asking
myself
here
or
also
we
set
up
the
experiment.
It's
okay.
B
We
have
the
problem
of,
we
are
doing
monitoring
and
we
rely
on
the
fact
that
okay,
we
might
even
have
artifacts
on
absolute
measurements,
with
making
the
differences
we
you
really
should
be
able
to
see
processes,
but
the
water,
the
adding
the
water
is
not
the
only
thing
that's
happening,
there's
also
erosion.
Typically
in
these
things,
like
you
can
imagine
that
these
bumps
in
the
beginning,
they
are
nice
and
smooth
and
a
machine
makes
them,
but
then
there's
rain
wind.
So
this
micro
topography
will
degrade
over
time.
F
B
B
Yeah
because
it's
a
human-made
soil.
Actually,
this
is
what
farmers
do
in
this
region,
because
it's
a
very
poor
soil
is
that
over
hundreds
of
year,
they've
added
manure
and
organic
material
from
the
forests
on
top
of
their
fields,
and
this
makes
such
a
soil
that,
together
with
some
kind
of
humidification
process
that
is
taking
place
in
these
real
actual,
so
is
there
so
initially.
B
B
Is
a
bit
similar
in
terms
of
accumulation
but
anyway,
and
then
the
the
last
thing
that
I
was
or
yeah
last
thought
that
I
have
before
we
started
these
loci
there's
also,
these
distinct
patterns
quite
different
that
we
want
to
recognize.
We
want
to
be
able
to
measure
an
infiltration
front
which
is
a
homogeneous
line,
going
the
untrue,
your
profile,
something
horizontal,
whereas
Rueter
uptake
is
something
very
local
and
a
bulb,
which
is
a
very
different
pattern
to
to
distinguish.
B
So
we
have
to
be
sure
that
the
array
that
we
use
is
also
sensitive
to
both
kind
of
structures
and
not
trying
to
deform
one
or
the
other.
So
that's
a
bit
the
context,
and
so
what
we
did
is
like
study
into
world
like
we
started
with
a
virtual
experience,
as
you
guys
often
do
so.
I
wanted
to
use
actually
for
modeling
to
have
a
better
idea
of
this
erosion
problem.
So
how
which
kind
of
artifact
do
we
have
to
expect
when
we
have
a
wrong
estimation
of
this
erosion
of
the
extent
of
the
erosion?
B
So
here
you
can
just
see
a
loop
of
like
simulation
of
the
hydrological
situation.
We
generate
some
typical
patterns,
we
convert
them
from
so
much
to
connectivity,
even
a
better
physical
relationship,
and
then
we
do
the
typical
loop
for
modeling
inverse
modeling
in
comparing
the
real
model
and
the
inverted
model.
So
that's
for
the
question
of
the
artifacts
and
then
we
had
a
field
experiment
where
we
really
wanted
to
look
mainly
qualitatively
because
for
the
petrol
physics,
we
still
have
a
lot
of
work.
G
B
Where
we
can
monitor
this
infiltration,
that's
actually
something
feasible
in
reality,
and
what
do
these
patterns
look
like?
So
that's
the
same
kind
of
story,
but
there
we
have
real
deal.
So
we
start
a
bedside.
We
do
an
inversion
and
then
we
should
get
some
qualitative
information,
at
least
on
the
distribution
of
soil
moisture.
So
that's
what
I
did
here.
I
started
off
with
virtual
experiments,
so
I
use
a
lot
of
dumps.
Knowledge
don't
actually
have
to
generate
a
very
beautiful
mesh.
I
think
it's
almost
something.
B
I
put
my
wall,
so
you
see
this
typical
bumps
here.
Triangles
are
the
electrodes
locations
and,
as
you
can
see
here
were
really
small
scale.
So
if
you
would
project
the
electrodes
on
the
x-axis,
you
would
have
lists
of
around
six
point.
Six
centimeters
between
them.
Give
you
an
idea.
It's
a
bit
wider
on
the
top
two,
because
the
plant
is
there.
We
did
not
put
a
label
of
a
mr.
LaPlante,
of
course,
yeah.
We
are
using
1
or
alpha
and
I
pull
dipole
quadruples
around
first
300
measurements,
which
should
be
divided
by
2.
B
B
So
that's
probably
something
we
could
discuss
it
if
you
have
some
hints
or
tips
to
improve
still
the
results
so
simple,
regularization
in
exact
cause,
Newton,
updated
sensitivity,
weights
and
target
target
misfit
and
as
a
reference
model,
I
just
used
the
medium
of
the
apparent
resistivity
just
to
because
I
think
I
didn't
want
to
force
the
model
towards
what
we
know
already
in
terms
of
layering,
because
in
reality
we
don't
know
anything
so
I
just
put
I'm
a
genius
background
to
the
reference
model.
So.
B
For
the
transition
between
the
two
horizons-
so
that's
maybe
a
bit
of
it's
true.
This
is
also
something
we
probably
wouldn't
know
in
reality,
because
we
don't
know
exactly
where
second
horizon
would
be,
but
they
yeah.
We
know
that
hydrologically
there's
always
a
lot
going
on
boundaries
between
two
soil
horizons
and
also.
B
F
B
F
B
Something
I
could
change
I'm
happy
to
hear
these
things
like
I.
Just
did
it
also
yeah
to
be
sure
that
we
cover
and
easy
arrange
things
happening
there,
but
yeah
might
be
good
to
at
least
try
without
refining
it
to
see
ya,
because
it
artificially
increases
the
sensitivity
to
a
bit
there.
Of
course,
if
you
find
ourselves
it's
one
of
the
things,
I
could
look
at
them.
Thank
you,
yeah
yeah,
so
doesn't
just
look
at
the
results.
B
I
just
took
out
the
most
extreme
erosion
like
we
take
off
six
centimeters
of
the
height
and
the
normal
case.
What
we
did
is
we
similar
forward
simulated
all
cases
with
the
correct
micro
topography.
So
that's
what
you
see
in
the
first
column
like
they
would,
there
would
be
an
irrigation
event,
so
is
infiltration
with
twenty
centimeters
one
with
ridges
of
14
centimeter
side.
Only
so
that's
the
real
model
on
the
left
and
then
the
next
column.
Is
we
invert
that
with
the
correct
macro
topography?
So
what
do
we
get
them?
B
You
see
already
that
it's
hard
to
reproduce
this
transition
between
the
horizons,
for
example,
and
then
in
the
last
column
we
see
the
image
where
I
ignored
effects
that
there
has
been
erosion,
so
I
invert,
as
if
my
electrodes
are
still
in
the
same
place
as
if
I
I,
don't
know
that
there
has
been
any
erosion.
So
I
keep
my
initial
macro
to
apocrypha
and
I
invert,
the
data
of
the
aerodromes
micro
topography.
B
Then
you
can
see
that
yeah
there's
some
artifacts
that
are
being
produced
that
mainly
everything
has
changed
in
the
most
sensitive
parts
of
the
yachty
array,
of
course,
which
is
close
to
the
surface
and
in
the
ridges
and
the
same
I.
Do
the
same
for
the
second
scenario,
where
I
was
routed
or
uptake
so
a
different
type
of
pattern,
and
there
yeah,
you
can
see
the
same
effect,
but
it's
even
more
strongly
there
because
yeah,
you
see,
it's
mainly
concentrating
everything
in
the
ridges
at
that
point.
So
this
is
the
most
extreme
case.
B
D
B
Could
actually
do
that,
but
the
point
is
I:
don't
have
a
lot
of
information
about
constraining
that
I
mentioned
so
I
think
in
this
case
it's
not
very
useful
when
we
would
do
transversal,
transacts
like
on
the
rich
of
course,
it's
necessary
to
go
to
treaty
okay,
so
just
to
give
you
an
idea
of
the
difference
between
if
we
start
using
this
wrong-
micro
typography,
we
get
much
higher
data
misfits
of
course
too.
So
that's
maybe
like
something
also
to
take
with
a
take-home
message.
Okay,
we
have
some
artifact,
but
it's
also
visible.
B
H
B
J
B
B
B
E
F
B
Where
are
you
with
them?
What
I
don't
like
when
I
think
that
the
problem
is
that
the
sensitivity
of
the
array
is
very
superficial,
so
the
problem
for
me:
it's
a
problem
that
the
the
horizons
are
disappearing
because
you
start
you
could
start
interpreting
these
as,
for
example,
drying
bulbs,
or
something
like
that
or
wetting
bulbs,
instead
of
it's
just
the
value
of
the
rise,
and
actually
that
you
see
there.
So
you
get
strange
structures.
B
B
If
you
see
this
here,
this
is
the
model,
and
here
you
you
have
these
green
bulbs,
which
are
remaining
actually
from
this
background
values
of
the
horizon.
You
see,
so
you
this
seems
like
something
which
has
to
do
with
the
bridge,
but
actually
is
just
related
to
the
value
of
this
horizon,
but
it's
disappearing
because
the
inversion
is
tearing
the
the
resistive
values
of
ridges
down.
F
B
What
is
good,
for
example,
is
that
we
seem
to
capture
this
infiltration
fronts
quite
nicely,
and
the
thing
that
I
don't
like
is
indeed
that
it's,
but
that's
probably
the
array,
it's
nothing.
It's
not
the
inversion
theory,
which
is
not
captured,
not
going
deep
enough
to
really
be
sensitive
to
both
arising
so
his
horizons.
These
things
are
very
independent
on
the
background
and
reference
for
all
the
values
and
what
is
well
I
am
now
trying
several
so
to
see.
B
What's
the
impact
of
the
way
it
reaches,
but
you
see
that
if
you
use
one
which
is
close
to
the
values
on
topic,
its
smearing
everything
out
to
the
bottom
and
if
you
use
one
which
is
further
away
from
that,
it's
it's
not
doing
that.
So
it
can
have
quite
an
impact
actually
from
half
a
meter
downwards.
So.
B
A
A
B
H
B
No
you're
a
bit
yeah,
it's
more
or
less
up
till
stick
your
point,
six
that
you
gets
sensitivity
and
the
rest
is
much
less
yeah,
but-
and
that's
here
just
the
same-
but
if
you
do
it
in
a
one
deal
with
which
just
and
the
profiles
just
to
show
that
the
stars
are
actually
the
one
which
are
inverted
with
one
micro
micro
typography,
and
then
you
see
that
mainly
deviates
at
the
surface,
always
just
the
same
message.
We
are
actually
creating
artifacts,
very
close
to
the
surface
with
this
kind
of
problem.
B
B
So
you
have
like
about
nine
measurements
over
a
few
days,
yeah
to
get
some
idea
of
how
things
are
changing
over
time
and
we
have
some
standards
also
in
there
the
Stars
and
is
drawing
our
sensors,
so
I
wash
your
sensors
and
temperature
sensors
to
get
to
get
to
the
Petra
physics
section.
To
give
you
an
idea.
What
we
did
here
is
now.
B
This
is
the
climates
before,
and
you
can
see
that
apart,
starting
in
the
end
of
whoops
here
in
the
end
of
May,
we
didn't
have
any
red
anymore,
so
it
was
a
very
dry
season
and
then
this
these
green
bars
are
the
irrigation.
So
the
farmer
gets
a
lot
this
year
and
what
we
measured
is
one
of
these
events.
So
you
see
here
this
one
we
had
even
more
than
three
year
the
RT
measurements
I
have
to
update
this
figure.
B
B
So
this
is
the
the
ERT
inversion.
I
try
to
different
error
levels
and
different
settings
and
I
mean
again.
You
strongly
have
this
smearing
down
roots
of
the
structures
in
the
top,
so
I
don't
know.
Is
there
something
else
that
I
could
try
there?
But
let's
discuss
that
later
because
it
seemed
it's
always
important
to
look
at
this
time.
Lapse.
Result
and
that's
what
you
see
here
so
here
you
see
basically
the
difference
between
the
first
inversion
and
then
every
time
step
divided
by
the
first
one
to
a
relative
change.
B
So
you
can
see
that
where
the
water
is
coming
in
these
curves
that
we
really
are
talking
about
big
changes,
I
it's
more
than
15
times
the
initial
like
the
initial
connectivity,
so
that
the
blue
values-
and
this
way
you
can
it's
not
so
visible
here
so
I
should
probably
use
another
color
scale
yeah.
We
do
see
now
that
there
is
mainly
infiltration
in
the
first.
If
we
make
these
difference,
there
is
no
artifacts
pop
popping
up
one
time
and
then
disappearing
again,
that's
also
nice
to
see.
B
So
that
means
that
if
we
have
artifacts
to
truly
inversion,
they
are
constant
over
time
and
so
they're
not
like
disturbing
our
process
understanding,
and
we
do
see
that,
for
example,
here
in
the
beginning,
we
also
have
some
ridges
where
water
is
arriving,
it's
not
only
in
the
first,
so
we
probably
have
a
mix
between
the
processes
that
I've
been
describing
it
in
the
beginning.
Well,
anyway,
so
qualitatively,
that's
what
we
see
we
have.
The
strong
connectivity
increases
that
it
gradually
decreases
over
time,
which
is
also
logical.
B
What
is
also
nice
that
we
don't
seem
to
have
this
kind
of
drainage
towards
the
bottom
of
the
profile,
so
it
does
say
like
there's
a
strong
increase,
but
it's
not
going
through
entirely
so
might
be
later,
it's
ponding
or
like
a
big
increase
of
water,
but
then
it
gradually
dissipates
towards
the
rest
of
the
profile
through
capillary
force
and
then
the
ridges
to
let
you
have
some
thing.
There's
and
for
me,
conclusions
that
disappoints
are
that
we
began
with
timeless.
Ert
still
reveal
some
patterns
that
are
relevant
to
the
process.
B
So
the
conclusion
that
I
just
made
that
is
first
mainly
strange,
seems
to
be
Twitter,
but
then
there
is
one
thing
that
I
still
wanted
to
say:
here's,
because
that's
that
will
be
the
next
thing
for
me
is
connectivity,
is,
of
course
not
equal
to
so
much
here.
We
have
a
nonlinear
relationship
between
both
so
to
really
look
at
patterns.
I
should
convert
to
so
much
term,
because
some
connectivity
changes
will
not
be
very
relevant
in
terms
of
service
check
and
others
will
be
there's
an
impact
of
these
huge
potatoes
also
because
they
grow.
B
Three
days
until,
but
if
you
study
an
entire
growing
season,
then
I'm
sure
that
this
will
impact
also
the
connectivity,
and
then
you
have
some
idea
how
to
change
your
physics,
of
course,
for
this
kind
of
living
system
and
temperature
I'm
British
permits.
Yes,
all
right.
This
is
probably
a
naive
widows.
A
conductor
or
resist
well
I
have
entire
project
at
the
moment,
which
is
called
erode
on
the
actual
signature
of
plants,
and
it's
not
an
easy
answer.
B
I
would
think
with
the
potato
depends
on
the
type
of
measurements
you're
doing,
but
with
ERT,
it's
probably
not
necessarily
acting
as
a
conductor,
because
it's
the
cells
are
protected
by
the
cell
wall
and
you're
not
going
to
get
easy
current
truth.
But
if
you
would
put
current
on
the
plant,
for
example
and
measure
in
the
soil,
you
might
have
a
very
big
conductor,
because
then
you
get
like
certain
vessels
that
are
so.
It
is
quite
complicated
to
answer
to
that
question
very.
A
Nice
yeah,
it's.
B
Worth
studying,
probably
yeah
and
yeah
the
temperatures
for
me
an
easy
one.
Probably
we
should
have
a
look
at
F.
We
often
just
use
a
temperature
dissipation
model
in
soil
to
at
least
get
an
idea
of
game.
It
could
be
the
impact
because
the
water
was
in
summer,
really
ten
degrees
cooler
than
the
soil
equation,
so
that
could
start
having
impacts
on
the
petro
physics.
Of
course,
if
we
want
to
get
to
soil
moisture
so
question.
D
B
B
B
B
And
that's
why
the
suttas
so
much
answers
are
important.
Then
you
can.
If
you
have
the
changes,
you
can
just
use
them
to
calibrate
yeah,
so
questions
I
have,
for
you
is
mainly
yeah
I'm
still
working
on
this
impact
of
the
reference
model.
Stuff
like
that.
So
looking
with
that
sensitivity
and
how
different
backgrounds
models
are
yeah
affecting
this
sensitivity,
distribution,
but
then
questions
I
add
to
you
mainly
was:
are
there
anything
are
things
you
see
in
the
inversions
now
there
are
obvious
problems
that
I'm
not
taking
like.
B
B
They
have
the
real
experiments
like.
How
can
we
link
somehow
this
knowledge
that
we
have
on
typical
artifacts
to
expects
like
I,
don't
know
one
quantitatively?
How
can
we
do
that
and
now
I've
like
an
idea
where
there
appear
visually,
but
I
would
like
to
use
both
together
in
one
story,
you
really
like
to
show
knowledge
here:
I
want
to
take
that
into
account
when
I'm
evaluating
my
reading
versions.
So
if
you
have
some
ideas
for
this
I'm
also
curious.
K
B
Like
the
the
link
between
the
synthetic
and
the
real
example-
yes
yeah,
so
for
me,
I
now
have
like
this
visualization
spatially
of
artifacts.
That
I
can
expect
when
I
have
a
wrong
micro,
topography,
estimate
and
I
have
a
reversion
of
a
really
good
case,
where
I
only
have
the
initial
macro
topography.
So
would
there
be
a
way
to
connects
like
the
knowledge
we
have
on
artifacts,
to
expect
to
like
and
advice
the
patterns
that
I
have
in
the
real
image.
B
A
B
They're
done
some
thirteen
centimeters
deep,
and
so
what
we
do
now,
I'm,
not
modeling
the
electrodes
and
which
would
be
another
I
cannot
do
that
in
this
paper.
Let's
say,
but
it
would
be
something
to
study.
Is
that
now
I'm
using
a
point
electrodes,
which
is
at
60%
of
the
Buried
length
of
the
electrodes?
B
So
that's
what
you
see
there
in
the
drawing
and
if
it's
eroding
and
taking
in
turn
into
a
carton
soils
leaving
to
change
that
point,
like
imagine,
you
have
ten
centimeters
electrodes
in
the
soil
first
and
then
it's
starting
to
erosion
2-centimeter
leaves
then
I
put
the
point
at
60
centimetres
of
60%
of
the
remaining
buried
lengths.
That's
what
you
see
there
with
these
dots
there's
a
wearing
tomorrow,
nothing
happens,
but
the
soil
is
removed.
Then
these
dot
changes.
A
G
F
B
From
a
study
that
master
students
did,
he
did
a
lot
of
different
arrays
before
he
did
the
the
thing
with
the
erosion
he
tested
on
this
micro
topography,
different
types
of
race,
and
this,
this
combination
gives
the
best
gave
the
best
coverage
and
best
sensitivity
for
the
processes
that
we
want
to
measure.
So
it's
even
yeah
he
tested
more
than
than
just
these
two.
So
we
have
here
to
think.
B
Scale,
this
top
so
for
a
farmer
this
is
not
fitted,
I
mean
for
an
experiment.
You
could
scale
this
up,
and
but
you
need
this
resolution
because
I
think
already.
Our
resolution
is
limited
because
of
the
fact
that
we
have
this
4
inch.
Then
you
come
to
the
electrodes
anywhere.
But
if
you
want
to
see
change
in
soil
moisture
at
roots
linked
to
root,
folder
uptake,
you
have
to
have
this
resolution
serve.
But
for
me
that's
not
something
a
farm
do
like
that's!
Why
we
scale
up
in
this
project.
F
B
Yeah,
so
I
don't
want
to
take
too
much
of
your
time.
Tell
me:
did
you
actually
add
points
yeah,
I
ever
think
5%
noise
should
look
causing
noise
and,
like
your
scale,
Twitter
resistance
like
a
a
relative
component,
so
the
noise,
it's
a
relative
noise
with
the
right
stage
of
your
voltages,
yeah
and
just
add
a
5%
mbb
you
for.
F
G
J
B
B
J
A
D
J
B
A
A
If
not
the
last
item
that
I
wanted
to
chat
about
briefly
was
the
simulation
pull
request.
It's
still
very
much
a
work
in
progress,
but
I've
left
some
questions
and
things
in
there.
It
would
really
appreciate
input
on
because
this,
like
what
we're
looking
at
doing,
is
this
refactoring,
like
the
problem
at
you
know,
survey
so
you're
no
longer
well
will
no
longer
have
a
problem
class
at
all.
We'll
just
have
a
simulation
and
the
simulation
Cape
the
survey
and
a
mesh,
and
it
has
the
at
a
bare
minimum.
A
It
would
need
to
have
a
deep
red,
a
Jay
Beck
and
a
Katie
Boeck,
and
that's
it
and
so
that'll.
Let
us
actually
plug
in
other
codes
much
more
easily.
So
something
like
you
know
my
mods
and
example.
All
we
have
out
of
there
is
a
deep
red,
and
then
we
can
estimate
Jay,
Beck
and
JT
back
just
by
doing
like
differencing
on
a
different
angle,
and
so
then
we
should
be
able
to
immediately
plug
that
in
and
invert
and
I
think.
A
This
will
also
get
rid
of
the
naming
complications
that
we've
had
with
just
referring
to
problem
throughout
simple,
and
so,
as
I've
been
doing.
This
been
looking
at
renaming
some
things
because,
like
this
is
a
non
backwards,
compatible
change,
and
so
it's
like,
where
are
you
gonna,
have
to
update
a
number
of
things,
and
so
in
case
might
as
well
just
like
going
and
update
lots
of
things.
I
guess
trying
to
make
it
like
as
backwards-compatible
as
possible.
So
are
people
still
there.
I've
lost
video
for
everyone.