Agronomy,
Год журнала:
2024,
Номер
14(12), С. 3039 - 3039
Опубликована: Дек. 19, 2024
The
accurate
prediction
of
the
spatial
variability
for
soil
water
content
(SWC)
in
farmland
is
essential
resource
management
and
sustainable
agricultural
development.
However,
natural
factors
introduce
uncertainty
result
poor
alignment
when
predicting
SWC,
leading
to
low
accuracy.
To
address
this,
this
study
introduced
a
novel
indicator:
landscape
indices.
These
indices
include
largest
patch
index
(LPI),
edge
density
(ED),
aggregation
(AI),
cohesion
(COH),
contagion
(CON),
division
(DIV),
percentage
like
adjacencies
(PLA),
Shannon
evenness
(SHEI),
diversity
(SHDI).
A
Bayesian
optimization–deep
forest
(BO–DF)
model
was
developed
leverage
these
SWC.
Statistical
analysis
revealed
that
exhibited
skewed
distributions
weak
linear
correlations
with
SWC
(r
<
0.2).
Despite
incorporating
variables
into
BO–DF
significantly
improved
accuracy,
R2
increasing
by
35.85%.
This
demonstrated
robust
nonlinear
fitting
capability
Spatial
mapping
using
indicated
high-value
areas
were
predominantly
located
eastern
southern
regions
Yellow
River
Delta
China.
Furthermore,
SHapley
additive
explanation
(SHAP)
highlighted
key
drivers
findings
underscore
potential
as
valuable
prediction,
supporting
regional
strategies
Natural Hazards Research,
Год журнала:
2024,
Номер
unknown
Опубликована: Май 1, 2024
Gully
erosion
is
a
significant
global
threat
to
socioeconomic
and
environmental
sustainability,
making
it
widespread
natural
hazard.
Developing
spatial
models
for
gully
crucial
local
governance
effectively
implement
mitigation
measures
promote
regional
development.
This
study
applied
two
machine
learning
(ML)
models,
RF
XGB,
alongside
an
AHP-based
multi-criteria
decision
method
FR
bivariate
statistics,
assess
susceptibility
(GES)
in
the
Kangsabati
River
basin
eastern
India's
Chotonagpur
plateau
fringe.
A
GIS
database
was
created,
incorporating
recorded
incidents
20
conditioning
variables,
which
were
evaluated
multicollinearity.
These
variables
served
as
predictive
factors
assessing
presence
area.
The
models'
performance
using
metrics
such
RMSE,
MAE,
specificity,
sensitivity,
accuracy.
XGB
model
outperformed
others,
achieving
accuracy
of
90.22%.
found
that
approximately
6.56%
catchment
highly
susceptible
erosion,
with
12.39%
moderately
81.05%
not
susceptible.
had
highest
ROC
value
85.5
during
testing,
indicating
its
superiority
over
(ROC
=
81.7),
AHP
79.8),
83.8)
models.
findings
highlight
model's
efficacy
potential
large-scale
GES
mapping.
Biological Journal of the Linnean Society,
Год журнала:
2025,
Номер
145(2)
Опубликована: Июнь 1, 2025
Abstract
Malania
oleifera
is
an
endemic
endangered
tree
species.
The
effects
of
future
climate
change
on
the
distribution
pattern
suitable
habitat
for
M.
and
its
degree
fragmentation
are
unclear.
aim
this
research
was
to
investigate
impact
provide
insights
into
conservation
strategies
in
China.
results
showed
that
influenced
primarily
by
precipitation
driest
quarter,
isothermality,
mean
ultraviolet-B
lowest
month.
Areas
high
suitability
concentrated
Guangnan
Funing
counties
Yunnan
Province,
alongside
select
Guangxi
Guizhou
provinces.
It
worth
noting
habitats
anticipated
diminish
2050s
2070s.
A
reduction
largest
patch
index
cohesion
indicates
areas
becoming
less
contiguous.
An
increase
number
patches,
density,
landscape
division,
splitting
increasingly
fragmented.
Our
study
environmental
factors
pivotal
oleifera,
with
posing
significant
threats
habitat.
Hydrological Processes,
Год журнала:
2025,
Номер
39(6)
Опубликована: Июнь 1, 2025
ABSTRACT
Understanding
and
quantifying
the
spatiotemporal
variations
in
regional
soil
erosion
sediment
yield,
as
well
identifying
driving
factors,
are
crucial
for
managing
land
resources
addressing
environmental
issues
induced
by
erosion.
However,
critical
challenges
persist
semi‐arid
basins
due
to
interplay
of
wind‐water
processes
nonlinear
responses
coupled
climatic‐anthropogenic
drivers.
This
study
addresses
these
integrating
InVEST
(v3.14.1)
Sediment
Delivery
Ratio
(SDR)
model
with
geographical
detector
method.
We
analysed
spatial
temporal
changes
output
Xiliugou
basin
from
1990
2020
clarified
factors
behind
yield
SDR.
The
results
revealed
significant
annual
fluctuations
2020,
ranging
0.01
×
10
4
t
2011
1480
1998,
a
mean
(191.9
±
354.9)
t.
In
1990,
was
predominantly
distributed
upper
slopes
channels,
but
specific
decreased
substantially
after
2000.
Annual
rates
2000,
2010
were
99.98
,
30.51
44.62
45.30
t,
respectively.
Slight
dominated
regime,
accounting
67.7%–97.4%
watershed
area,
post‐2000
values
exceeding
90%.
Slope‐driven
heterogeneity
most
pronounced
factor
influencing
distribution.
Interactions
between
slope
(a
topographic
factor)
climatic
anthropogenic
drivers
significantly
amplified
their
impacts
on
patterns.
decline
load,
primarily
driven
reduced
hyperconcentrated
flows,
vegetation
cover
changes,
sediment‐trapping
dam
constructions,
identified
main
contributor
SDR
basin.
