Geosciences,
Год журнала:
2024,
Номер
15(1), С. 2 - 2
Опубликована: Дек. 26, 2024
This
study
aims
to
establish
a
scientific
and
methodological
basis
for
predicting
shoreline
positions
using
modern
data
analysis
machine
learning
techniques.
The
focus
area
is
5
km
section
of
the
Ural
coast
along
Baydaratskaya
Bay
in
Kara
Sea.
region
was
selected
due
its
diverse
geomorphological
features,
varied
lithological
composition,
significant
presence
permafrost
processes,
all
contributing
complex
patterns
change.
Applying
advanced
methods,
including
correlation
factor
analysis,
enables
identification
natural
signs
that
highlight
areas
active
coastal
retreat.
These
insights
are
valuable
arctic
development
planning,
as
they
help
recognize
zones
at
highest
risk
transformation.
erosion
process
can
be
conceptualized
comprising
two
primary
components
construct
predictive
model
first
random
variable
encapsulates
effects
local
structural
changes
coastline
alongside
fluctuations
climatic
conditions.
component
statistically
characterized
define
confidence
interval
variability.
second
represents
systematic
shift,
which
reflects
regular
average
over
time.
more
suited
modeling.
Thus,
information
processing
methods
allow
us
move
from
descriptive
numerical
assessments
dynamics
processes.
goal
ultimately
support
responsible
sustainable
highly
sensitive
region.
Remote Sensing,
Год журнала:
2025,
Номер
17(4), С. 594 - 594
Опубликована: Фев. 10, 2025
This
study
explores
the
potential
of
repurposing
historical
aerial
photographs
to
produce
high-accuracy
digital
surface
models
(DSMs)
at
regional
scales.
A
novel
methodology
is
introduced,
incorporating
road
points
for
quality
control
and
refinement
enhance
precision
comparability
multitemporal
DSMs.
The
method
consists
two
phases.
first
photogrammetric
phase,
where
DSMs
are
generated
using
structure
from
motion
(SfM)
techniques.
second
which
uses
a
large
number
(millions)
extracted
centrelines
evaluate
altimetric
residuals—defined
as
differences
between
reference
DSM.
These
filtered
ensure
that
they
represent
stable
positions.
analysis
shows
initial
residuals
exhibit
geographical
trends,
rather
than
random
behaviour,
removed
after
refinement.
An
application
example
covering
whole
coast
Valencian
region
(Eastern
Spain,
518
km
coastline)
obtention
series
composed
six
achieves
levels
accuracy
(0.15–0.20
m)
comparable
modern
LiDAR
techniques,
offering
cost-effective
alternative
three-dimensional
characterisation.
foredune
coastal
environment
demonstrated
method’s
effectiveness
in
quantifying
sand
volumetric
changes
through
comparison
with
achieved
crucial
establishing
precise
sedimentary
balances,
essential
management.
At
same
time,
this
significant
its
other
dynamic
landscapes,
well
urban
or
agricultural
monitoring.
