Earth system science data,
Journal Year:
2024,
Volume and Issue:
16(8), P. 3795 - 3819
Published: Aug. 27, 2024
Abstract.
Land
surface
temperature
(LST)
serves
as
a
crucial
variable
in
characterizing
climatological,
agricultural,
ecological,
and
hydrological
processes.
Thermal
infrared
(TIR)
remote
sensing
provides
high
temporal
spatial
resolutions
for
obtaining
LST
information.
Nevertheless,
TIR-based
satellite
products
frequently
exhibit
missing
values
due
to
cloud
interference.
Prior
research
on
estimating
all-weather
instantaneous
has
predominantly
concentrated
regional
or
continental
scales.
This
study
involved
generating
global
daily
mean
product
spanning
from
2000
2020
using
XGBoost.
Multisource
data,
including
Moderate-Resolution
Imaging
Spectroradiometer
(MODIS)
top-of-atmosphere
(TOA)
observations,
radiation
products,
reanalysis
were
employed.
Validation
an
independent
dataset
of
77
individual
stations
demonstrated
the
accuracy
our
yielding
root
squared
errors
(RMSEs)
2.787
K
(instantaneous)
2.175
(daily).
The
RMSE
clear-sky
conditions
was
2.614
product,
which
is
slightly
lower
than
cloudy-sky
2.931
K.
Our
higher
compared
MODIS
official
(instantaneous
=
3.583
K;
3.105
K)
land
component
fifth
generation
European
ReAnalysis
(ERA5-Land)
4.048
2.988
K).
Significant
improvements
are
observed
notably
at
latitudes,
product.
monthly
scale,
first
day
2010
can
be
freely
downloaded
https://doi.org/10.5281/zenodo.4292068
(Li
et
al.,
2024),
complete
will
available
https://glass-product.bnu.edu.cn/
(last
access:
22
August
2024).
Remote Sensing,
Journal Year:
2025,
Volume and Issue:
17(5), P. 837 - 837
Published: Feb. 27, 2025
Soil
moisture
(SM)
monitoring
in
farmland
at
a
regional
scale
is
crucial
for
precision
irrigation
management
and
ensuring
food
security.
However,
existing
methods
SM
estimation
encounter
significant
challenges
related
to
accuracy,
generalizability,
automation.
This
study
proposes
an
integrated
data
fusion
method
systematically
assess
the
potential
of
three
automated
machine
learning
(AutoML)
frameworks—tree-based
pipeline
optimization
tool
(TPOT),
AutoGluon,
H2O
AutoML—in
retrieving
SM.
To
evaluate
impact
input
variables
on
six
scenarios
were
designed:
multispectral
(MS),
thermal
infrared
(TIR),
MS
combined
with
TIR,
auxiliary
data,
TIR
comprehensive
combination
MS,
data.
The
research
was
conducted
winter
wheat
cultivation
area
within
People’s
Victory
Canal
Irrigation
Area,
focusing
0–40
cm
soil
layer.
results
revealed
that
scenario
incorporating
all
types
(MS
+
auxiliary)
achieved
highest
retrieval
accuracy.
Under
this
scenario,
AutoML
frameworks
demonstrated
optimal
performance.
AutoGluon
superior
performance
most
scenarios,
particularly
excelling
scenario.
It
accuracy
Pearson
correlation
coefficient
(R)
value
0.822,
root
mean
square
error
(RMSE)
0.038
cm3/cm3,
relative
(RRMSE)
16.46%.
underscores
critical
role
strategies
enhancing
highlights
advantages
regional-scale
retrieval.
findings
offer
robust
technical
foundation
theoretical
guidance
advancing
efficient
monitoring.
Remote Sensing,
Journal Year:
2025,
Volume and Issue:
17(5), P. 892 - 892
Published: March 3, 2025
Compared
with
atmospheric
urban
heat
islands,
surface
islands
(SUHIs)
are
easily
monitored
by
the
thermal
sensors
on
satellites
and
have
a
more
stable
spatial
pattern
resembling
built-up
lands
across
single
cities,
large
metropolitans,
agglomerations;
hence,
they
gaining
attention
from
scholars
planners
worldwide
in
search
for
reasonable
patterns
scales
to
guide
future
development.
Traditional
urban–rural
dichotomies,
being
sensitive
representative
rural
areas
diurnal
seasonal
variations
land
temperature
(LST),
obtain
inflated
varying
SUHI
footprints
of
approximately
1.0–6.5
times
size
different
satellite-retrieved
LST
datasets
many
cities
metropolitan
areas,
which
not
conducive
developing
strategies
mitigate
SUHIs.
Taking
Yangtze
River
Delta
agglomeration
China
as
an
example,
we
proposed
improved
structural
similarity
index
quantify
SUHIs
multiple
at
annual
interval.
We
identified
gridded
anomalies
(LSTAs)
related
urbanization
adopting
random
forest
models
climate,
urbanization,
geographical,
biophysical,
topographical
parameters.
Using
LSTA
cycle
grid
point
relative
reference
terms
average
values,
variances,
shapes
characterize
SUHIs,
cross-validated
~1.06–2.45
×
104
km2
smaller
than
clear
transition
zones
between
zone
were
obtained
2000–2022.
Hence,
can
balance
urbanization’s
benefits
adverse
effects
enhancing
design.
Considering
that
rapidly
transformed
into
ratio
extent
increasing
0.43
0.62
during
2000–2022,
should
also
take
measures
prevent
rapid
expansion
high-density
ISA
density
above
65%
