Remote Sensing,
Год журнала:
2024,
Номер
17(1), С. 34 - 34
Опубликована: Дек. 26, 2024
In
the
current
work,
authors
reviewed
latest
research
results
in
landslide
susceptibility
mapping
(LSM)
using
artificial
intelligence
(AI)
methods.
Based
on
an
overall
review
of
collected
publications,
was
classified
into
four
sections
based
their
complexity:
single-model
approaches,
enhanced
models
with
optimization,
ensemble
models,
and
hybrid
models.
Each
category
offers
distinct
advantages
is
suited
to
specific
geographic
data
conditions,
enabling
selection
optimal
model
type
complexity
requirements
task.
Among
random
forest
(RF),
support
vector
machine
(SVM),
convolutional
neural
network
(CNN),
multilayer
perception
(MLP)
are
used
as
baseline
compare
any
new
introduced
develop
LSM.
Moreover,
compared
previous
works,
number
LSM
conditioning
factors
AI
significantly
increased,
up
122
factors.
Their
relation
illustrated
Sankey
diagram,
while
a
radar
chart
further
visualize
dataset
size
per
work
for
comparative
purposes.
main
part
findings
summarized
table
form,
where
reader
can
find
relations
between
factors,
size,
applied
accuracy
predicting
selected
geographical
locations.
terms
regions,
Asia
leading
application
generate
LSM,
such
regions
dense
populations
falling
higher
risk
categories,
there
more
ongoing
activities,
modern
This
trend
underscores
increased
use
disaster
management,
implications
improving
practical
applications,
early
warning
systems
informing
policy
decisions
aimed
at
reduction
vulnerable
areas.
IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing,
Год журнала:
2023,
Номер
16, С. 1 - 18
Опубликована: Янв. 1, 2023
Construction
activities
of
accelerated
urbanization
in
Shenzhen
have
increased
the
landslide
risk
area,
which
has
intensified
potential
threat
to
human
and
natural
environment.
However,
landslides
is
poorly
evaluated.
In
this
paper,
a
evaluation
(LRE)
model
constructed
using
susceptibility
map
(LSM)
vulnerability.
experiment,
stacking
ensemble
learning
(SEL)
based
on
convolutional
neural
network
(CNN),
multilayer
perceptron
(MLP),
gated
recurrent
unit
(GRU)
support
vector
machine
regression
(SVR)
generate
LSM
by
topography,
geology,
engineering
activities,
time-series
precipitation
normalized
difference
vegetation
index
(NDVI).
Road
network,
building
distribution
density
annual
average
data
are
used
evaluate
vulnerability
entropy
weight
method.
study,
multiple
statistical
indicators
performance
model,
Interferometric
Synthetic
Aperture
Radar
(InSAR)
deformation
utilized
verify
LRE
results
Shenzhen.
The
show
that
SEL
method
more
refined
for
LSM,
with
best
overall
accuracy,
especially
receiver
operating
characteristic
curve
(ROC),
where
accuracy
improved
nearly
8%.
Shenzhen,
very
high,
moderate,
low
areas
account
0.283%,
0.451%,
0.859%,
36.890%
61.517%,
respectively.
most
high
InSAR
clear
concentrated
trend
large
rate.
Research
can
provide
technical
disaster
prevention
Geomatics Natural Hazards and Risk,
Год журнала:
2024,
Номер
15(1)
Опубликована: Авг. 21, 2024
Crafting
landslide
susceptibility
mapping
is
pivotal
for
the
effective
management
of
risks.
However,
influence
non-landslide
sample
selection
on
modeling
performance
assessment
models
remains
a
crucial
challenge
to
overcome.
This
article
employs
Huize
County
as
research
area
and
identifies
12
factors
that
exert
influence.
In
this
study,
we
utilized
Extreme
Gradient
Boosting
Random
Forest
algorithms,
four
methods
(Whole-area
random
method,
Buffer
Frequency
Ratio
Analysis
Hierarchy
Process)
were
employed
select
samples
constructing
model.
The
findings
revealed
model
derived
from
different
selections
exhibited
significant
variations,
obtained
using
buffer
zone
frequency
ratio
AHP
method
performed
better
than
full-area
Among
evaluated
models,
demonstrated
most
optimal
performance,
with
an
AUC
92.17%
XGBoost-AHP
91.64%
RF-AHP.
Based
SHapley
Additive
explanation
(SHAP),
main
variables
impacting
danger
in
elevation,
NDVI,
peak
seismic
acceleration.
study
provides
theoretical
support
assessments
interpretable
AI
research.
Geocarto International,
Год журнала:
2023,
Номер
38(1)
Опубликована: Окт. 23, 2023
This
study
makes
a
significant
contribution
to
the
field
of
groundwater
potential
mapping
(GWPM)
by
exploring
application
ensemble
learning
models
(ELMs),
specifically
boosting
(BEMs),
which
have
not
been
fully
utilized
in
GWPM.
By
employing
six
ELMs
(random
forest,
AdaBoost,
XGBoost,
CatBoost,
GBDT,
and
LightGBM),
along
with
Tree-structured
Parzen
Estimator
Luoning
County,
China,
this
identifies
key
indicators
(topographic
position
index,
distance
rivers,
topographic
wetness
index)
demonstrates
superior
model
performance
XGBoost
compared
other
ELMs.
Additionally,
correlation
analysis
confirms
accuracy
predicting
relationships
between
important
potentials.
Finally,
findings
provide
valuable
insights
for
sustainable
management
strategies
County
emphasize
need
further
exploration
ELMs,
development
comprehensive
evaluation
indicator
systems,
reduction
inconsistencies
predication
results
practical
research
support
future
management.
The Science of The Total Environment,
Год журнала:
2024,
Номер
921, С. 171152 - 171152
Опубликована: Фев. 24, 2024
With
urban
areas
projected
to
accommodate
68
%
of
the
global
population
by
2050,
imperative
for
inclusive,
safe,
and
sustainable
cities
becomes
paramount.
In
timeline
centers,
landslides
represent
one
most
destructive
phenomena,
involving
several
resources
allocation
with
private
public
investments,
sometimes
claiming
human
lives.
By
synergically
connecting
environmental,
planning,
configurational
spheres,
this
study
seeks
support
proactive
management
landslide
risk.
The
proposed
three-step
methodology
allowed
quantify
environmental
features
involved
in
occurrence,
evaluate
planning
framework
vulnerabilities,
suggest
alternative
configurations
that
experienced
landslides.
has
been
applied
case
a
tragic
Casamicciola
Terme
(Italy)
November
2022.
First,
stream
network
drainage
basin
corresponding
confluence
point
into
sea
have
calculated
(environmental
elaborations).
Subsequently,
these
elaborations
overlapped
runoff
mitigation
sediment
deposition
layers,
extracted
through
INVEST
software.
Secondly,
reconnaissance
local
superordinate
levels
realized,
deepen
tools
cogency
on
area,
contextually
deepening
constraints
characterize
it.
From
overlapping
two
steps,
free
risk
located.
Finally,
based
available
territorial
surface
(Sta)
cover
ratio
(Rct),
configuration
scenarios
proposed,
envisaging
relocation
buildings
landslide.
Results
show
originated
three
out
five
gullies.
Some
portions
are
still
under
high
very
hydrogeological
Contextually,
it
emerges
poor
attention
from
planners
framework.
Historic
settlement
an
Rct
33.64
%,
while
which
relocate
built
up
32,45
scenario
1
27,9
2.
resulted
useful
address
supporting
realization
scenarios.
We
expect
our
research
contribute
evolving
field
disaster
reduction,
providing
systematic
approach
manage
Abstract
High
Mountain
Asia
has
long
been
known
as
a
hotspot
for
landslide
risk,
and
studies
have
suggested
that
hazard
is
likely
to
increase
in
this
region
over
the
coming
decades.
Extreme
precipitation
may
become
more
frequent,
with
nonlinear
response
relative
increasing
global
temperatures.
However,
these
changes
are
geographically
varied.
This
article
maps
probable
hazard,
shown
by
indicator
(LHI)
derived
from
downscaled
temperature.
In
order
capture
of
slopes
extreme
precipitation,
simple
machine‐learning
model
was
trained
on
database
landslides
across
develop
regional
LHI.
applied
statistically
data
30
members
Seamless
System
Prediction
Earth
Research
large
ensembles
produce
range
possible
outcomes
under
Shared
Socioeconomic
Pathways
2‐4.5
5‐8.5.
The
LHI
reveals
will
most
parts
Asia.
Absolute
increases
be
highest
already
hazardous
areas
such
Central
Himalaya,
but
change
greatest
Tibetan
Plateau.
Even
regions
where
declines
year
2100,
it
prior
mid‐century
mark.
seasonal
cycle
occurrence
not
greatly
Although
substantial
uncertainty
remains
projections,
overall
direction
seems
reliable.
These
findings
highlight
importance
continued
analysis
inform
disaster
risk
reduction
strategies
stakeholders
Geoscience Frontiers,
Год журнала:
2023,
Номер
15(2), С. 101758 - 101758
Опубликована: Ноя. 22, 2023
With
the
development
of
landslide
monitoring
system,
many
attempts
have
been
made
to
predict
failure-time
utilizing
data
displacements.
Classical
models
(e.g.,
Verhulst,
GM
(1,1),
and
Saito
models)
that
consider
characteristics
displacement
determine
investigated
extensively.
In
practice,
is
continuously
implemented
with
data-set
updated,
meaning
predicted
life
expectancy
(i.e.,
lag
between
time
node
at
each
instant
conducting
prediction)
should
be
re-evaluated
time.
This
manner
termed
"dynamic
prediction".
However,
performances
classical
not
discussed
in
context
dynamic
prediction
yet.
this
study,
such
are
firstly,
disadvantages
then
reported,
incorporating
from
four
real
landslides.
Subsequently,
a
more
qualified
ensemble
model
proposed,
where
individual
integrated
by
machine
learning
(ML)-based
meta
model.
To
evaluate
quality
under
prediction,
novel
indicator
'discredit
index
(β)'
higher
value
β
indicates
lower
quality.
It
found
Verhulst
would
produce
results
significantly
β,
while
(1,1)
indicate
highest
mean
absolute
error
(MAE).
Meanwhile,
accurate
than
models.
Here,
performance
decision
tree
regression
(DTR)-based
best
among
various
ML-based
Land,
Год журнала:
2023,
Номер
12(8), С. 1558 - 1558
Опубликована: Авг. 6, 2023
Geological
disasters
refer
to
adverse
geological
phenomena
that
occur
under
the
influence
of
natural
or
human
factors
and
cause
damage
life
property.
Establishing
prevention
control
zones
based
on
disaster
risk
assessment
results
in
land
planning
management
is
crucial
for
ensuring
safe
regional
development.
In
recent
years,
there
has
been
an
increase
extreme
rainfall
events,
so
it
necessary
conduct
effective
hazard
assessments
different
conditions.
Based
first
national
survey
results,
this
paper
uses
analytic
hierarchy
process
(AHP)
combined
with
information
method
(IM)
construct
four
conditions,
namely,
10-year,
20-year,
50-year,
100-year
return
periods.
The
susceptibility,
hazard,
vulnerability,
Laoshan
District
eastern
China
are
evaluated,
established
evaluation
results.
show
that:
(1)
There
121
collapse
District,
generally
at
a
low
susceptibility
level.
(2)
A
positive
correlation
exists
between
hazards/risks.
With
condition
changing
from
10-year
period
period,
proportion
high-hazard
increased
20%
41%,
high-risk
31%
51%,
respectively.
Receiver
operating
characteristic
(ROC)
proved
accuracy
was
acceptable.
(3)
Key,
sub-key,
general
have
established,
corresponding
suggestions
proposed,
providing
reference
early
warning
other
regions.
Environmental Research Letters,
Год журнала:
2024,
Номер
19(12), С. 124016 - 124016
Опубликована: Окт. 23, 2024
Abstract
Mountainous
landslides
are
expected
to
worsen
due
environmental
changes,
yet
few
studies
have
quantified
their
future
risks.
To
address
this
gap,
we
conducted
a
comprehensive
analysis
of
the
eastern
Hindukush
region
Pakistan.
A
geospatial
database
was
developed,
and
logistic
regression
employed
evaluate
baseline
landslide
susceptibility
for
2020.
Using
latest
coupled
model
intercomparison
project
6
models
under
three
shared
socioeconomic
pathways
(SSPs)
cellular
automata-Markov
model,
projected
rainfall
land
use/land
cover
patterns
2040,
2070,
2100,
respectively.
Our
results
reveal
significant
changes
in
use
patterns,
particularly
long-term
(2070
2100).
Future
then
predicted
based
on
these
projections.
By
high-risk
areas
increase
substantially
all
SSP
scenarios,
with
largest
increases
observed
SSP5-8.5
(56.52%),
SSP2-4.5
(53.55%),
SSP1-2.6
(22.45%).
will
rise
by
43.08%
(SSP1-2.6),
40.88%
(SSP2-4.5),
12.60%
(SSP5-8.5).
However,
minimal
compared
baseline,
9.45%
1.69%
7.63%
These
findings
provide
crucial
insights
into
relationship
between
risks
support
development
climate
risk
mitigation,
planning,
disaster
management
strategies
mountainous
regions.