Remote Sensing,
Journal Year:
2023,
Volume and Issue:
15(17), P. 4159 - 4159
Published: Aug. 24, 2023
Landslides
are
devastating
natural
disasters
that
seriously
threaten
human
life
and
property.
Landslide
susceptibility
mapping
(LSM)
plays
a
key
role
in
landslide
hazard
management.
Machine
learning
(ML)
models
widely
used
LSM
but
suffer
from
limitations
such
as
overfitting
unreliable
accuracy.
To
improve
the
classification
performance
of
single
machine
model,
this
study
selects
logistic
regression
(LR),
support
vector
(SVM),
random
forest
(RF),
gradient
boosting
decision
tree
(GBDT),
proposes
novel
heterogeneous
ensemble
framework
based
on
Bayesian
optimization
(BO),
namely,
stratified
weighted
averaging
(SWA),
to
test
its
applicability
typical
area
Yanbian
Prefecture,
China.
Firstly,
dataset
consisting
1531
historical
landslides
was
collected
field
investigations
records,
spatial
database
containing
16
predisposing
factors
established.
The
divided
into
training
set
ratio
7:3.
results
showed
SWA
effectively
improved
Accuracy,
AUC,
robustness
model
compared
ML
model.
achieved
best
(Accuracy
=
91.39%
AUC
0.967).
verify
generalization
ability
SWA,
we
selected
published
datasets
Yanshan
country
Yongxin
China
for
testing.
also
performed
well,
with
an
0.871
0.860,
respectively.
As
indicated
by
shapely
values
(SVs),
Normalized
Difference
Vegetation
Index
(NDVI)
is
factor
has
greatest
impact
occurrence.
maps
obtained
will
provide
effective
reference
program
land
use
planning
disaster
prevention
mitigation
projects
Geoscience Frontiers,
Journal Year:
2023,
Volume and Issue:
15(2), P. 101773 - 101773
Published: Dec. 20, 2023
The
implementation
of
isolated
heterologous
monitoring
systems
for
spatially
distant
borehole
deployments
often
comes
with
substantial
equipment
costs,
which
can
limit
the
effectiveness
geohazard
mitigation
and
georisk
management
efforts.
To
address
this,
we
have
developed
a
novel
system
that
integrates
fiber
Bragg
grating
(FBG)
microelectromechanical
(MEMS)
techniques
to
capture
soil
moisture,
temperature,
sliding
resistance,
strain,
surface
tilt,
deep-seated
inclination.
This
enables
real-time,
simultaneous
data
acquisition
cross-validation
analyses,
offering
cost-effective
solution
critical
parameters
in
geohazard-prone
areas.
We
successfully
applied
this
integrated
Xinpu
landslide,
an
active
super-large
landslide
located
Three
Gorges
Reservoir
Area
(TGRA)
China.
resulting
strain
profile
confirmed
presence
two
shallow
secondary
surfaces
at
depths
approximately
7
m
12
m,
respectively,
addition
depth
∼28
m.
lower
was
activated
by
extreme
precipitation,
while
upper
one
primarily
driven
significant
changes
reservoir
water
levels
secondarily
triggered
concentrated
rainfalls.
Anti-slide
piles
remarkably
reinforced
moving
masses
but
failed
control
ones.
gap
between
pile
heads
amplified
rainwater
erosion
effect,
creating
preferential
channel
infiltration.
Multi-physical
measurements
revealed
mixture
seepage-driven
buoyancy-driven
behaviors
within
landslide.
study
offers
dual-source
multi-physical
paradigm
collaborative
multiple
crucial
boreholes
on
large-scale
contributes
evaluation
improvement
engineering
measures
similar
geological
settings.
Geomatics Natural Hazards and Risk,
Journal Year:
2024,
Volume and Issue:
15(1)
Published: July 17, 2024
Non-landslide
samples
influence
the
outcomes
of
landslide
susceptibility
assessment.
Existing
studies
did
not
fully
consider
equilibrium
between
and
non-landslide
in
similar
environments,
resulting
poor
reliability
This
study
proposed
a
optimization
method
with
constraint
disaster-pregnant
environment
similarity
to
construct
balanced
samples.
We
employed
heterogeneous
stacking
blending
ensemble
learning
models
generate
focused
on
Bailong
River
Basin
complex
frequent
landslides
as
area.
First,
we
extracted
12
influencing
factors
based
multiple
sources
analyzed
spatial
distribution
patterns
landslides.
Second,
constructed
environments
assessment
units
obtained
from
curvature
watershed
selected
an
equal
amount
both
every
different
environment.
Finally,
three
classic
neural
network
models,
namely
multilayer
perceptron,
convolutional
network,
gated
recurrent
unit
were
used
base
for
assess
susceptibility.
The
findings
suggested
that
results
optimized
more
reliable,
especially
improved
prediction
sample-sparse
regions.
this
demonstrated
highest
area
under
curve
0.88
testing
dataset,
outperforming
models.
issue
unreliable
regions
within
can
be
effectively
addressed
by
considering
sampling
environments.
Journal of King Saud University - Science,
Journal Year:
2024,
Volume and Issue:
36(5), P. 103174 - 103174
Published: March 20, 2024
Landslide
is
a
considerable
geomorphological
risk
in
terrain
systems
worldwide.
Advanced
techniques
present
unique
tool
for
predicting
landslide
susceptibility
with
unbiased
and
precise
outputs.
However,
the
application
of
this
to
analyze
eastern
Mediterranean
landscape
still
not
sufficiently
understood.
This
study
aimed
assess
implementation
three
machine
learning
(ML)
algorithms,
i.e.,
support
vector
(SVM),
random
forest
(RF)
extreme
gradient
boost
(XGBoost),
mapping
mountainous
area
western
Syria.
In
regard,
200
events
were
inventoried
from
historical
data,
aerial
images
conducted
fieldworks.
Sixteen
triggering
factors
selected
according
literature
geographical
features
(Monsoon
period).
The
receiver
operating
characteristic
(ROC)
outcomes
revealed
that
RF
achieved
better
performance
an
under
curve
(AUC)
0.96,
pursued
by
XGBoost
SVM
AUC
0.94
0.90,
respectively.
assessment
presents
essential
understanding
effective
ML
region
Mediterranean.
We
emphasized,
hence,
algorithm
has
most
robust
prediction
Moreover,
outputs
will
provide
local
decision-makers
insights
produce
regional
management
strategies
landslide,
especially
after
Syrian
war
phase.
