Information Systems Frontiers,
Journal Year:
2024,
Volume and Issue:
unknown
Published: June 24, 2024
Abstract
Forest
fires
have
far-reaching
consequences,
threatening
human
life,
economic
stability,
and
the
environment.
Understanding
dynamics
of
forest
is
crucial,
especially
in
high-incidence
regions.
In
this
work,
we
apply
deep
networks
to
simulate
spatiotemporal
progression
area
burnt
a
fire.
We
tackle
region
interpolation
problem
challenge
by
using
Conditional
Variational
Autoencoder
(CVAE)
model
generate
in-between
representations
on
evolution
area.
also
CVAE
forecast
fire
propagation,
estimating
at
distinct
horizons
propagation
stages.
evaluate
our
approach
against
other
established
techniques
real-world
data.
The
results
demonstrate
that
method
competitive
geometric
similarity
metrics
exhibits
superior
temporal
consistency
for
representation
generation.
context
forecasting,
achieves
scores
90%
99%
consistency.
These
findings
suggest
models
may
be
viable
alternative
modeling
2D
moving
regions
evolution.
International Journal of Wildland Fire,
Journal Year:
2024,
Volume and Issue:
33(3)
Published: March 18, 2024
Background
Fire
danger
rating
systems
are
used
daily
across
Australia
to
support
fire
management
operations
and
communications
the
general
public
regarding
potential
danger.
Aims
In
this
paper,
we
introduce
Australian
Danger
Rating
System
(AFDRS),
providing
a
short
historical
account
of
in
as
well
requirements
for
an
improved
forecast
system.
Methods
The
AFDRS
combines
nationally
consistent,
spatially
explicit
fuel
information
with
weather
advanced
behaviour
models
knowledge
produce
locally
relevant
ratings
potential.
Key
results
A
well-defined
framework
is
essential
categorising
defining
based
on
operational
response,
impact
observable
characteristics
incidents.
modular,
supporting
continuous
incremental
improvements
allowing
upgrades
components
response
new
science.
Conclusions
provides
method
estimate
best
available
models,
leading
potentially
significant
way
calculated,
interpreted.
Implications
was
implemented
2022,
most
change
forecasting
more
than
50
years.
Water Resources Research,
Journal Year:
2025,
Volume and Issue:
61(5)
Published: May 1, 2025
Abstract
This
paper
reviews
the
current
state
of
high‐resolution
remotely
sensed
soil
moisture
(SM)
and
evapotranspiration
(ET)
products
modeling,
coupling
relationship
between
SM
ET.
downscaling
approaches
for
satellite
passive
microwave
leverage
advances
in
artificial
intelligence
remote
sensing
using
visible,
near‐infrared,
thermal‐infrared,
synthetic
aperture
radar
sensors.
Remotely
ET
continues
to
advance
spatiotemporal
resolutions
from
MODIS
ECOSTRESS
Hydrosat
beyond.
These
enable
a
new
understanding
bio‐geo‐physical
controls
coupled
feedback
mechanisms
reflecting
land
cover
use
at
field
scale
(3–30
m,
daily).
Still,
state‐of‐the‐science
have
their
challenges
limitations,
which
we
detail
across
data,
retrieval
algorithms,
applications.
We
describe
roles
these
data
advancing
10
application
areas:
drought
assessment,
food
security,
precision
agriculture,
salinization,
wildfire
dust
monitoring,
flood
forecasting,
urban
water,
energy,
ecosystem
management,
ecohydrology,
biodiversity
conservation.
discuss
that
future
scientific
advancement
should
focus
on
developing
open‐access,
m),
sub‐daily
products,
enabling
evaluation
hydrological
processes
finer
scales
revolutionizing
societal
applications
data‐limited
regions
world,
especially
Global
South
socio‐economic
development.
