Using social media records to inform conservation planning
Conservation Biology,
Journal Year:
2023,
Volume and Issue:
38(1)
Published: Aug. 8, 2023
Citizen
science
plays
a
crucial
role
in
helping
monitor
biodiversity
and
inform
conservation.
With
the
widespread
use
of
smartphones,
many
people
share
information
on
social
media,
but
this
is
still
not
widely
used
Focusing
Bangladesh,
tropical
megadiverse
mega-populated
country,
we
examined
importance
media
records
conservation
decision-making.
We
collated
species
distribution
for
birds
butterflies
from
Facebook
Global
Biodiversity
Information
Facility
(GBIF),
grouped
them
into
GBIF-only
combined
GBIF
data,
investigated
differences
identifying
critical
areas.
Adding
data
to
improved
accuracy
systematic
planning
assessments
by
additional
important
areas
northwest,
southeast,
central
parts
extending
priority
4,000-10,000
km
Language: Английский
Large language models overcome the challenges of unstructured text data in ecology
bioRxiv (Cold Spring Harbor Laboratory),
Journal Year:
2024,
Volume and Issue:
unknown
Published: Jan. 26, 2024
Abstract
The
vast
volume
of
currently
available
unstructured
text
data,
such
as
research
papers,
news,
and
technical
report
shows
great
potential
for
ecological
research.
However,
manual
processing
data
is
labour-intensive,
posing
a
significant
challenge.
In
this
study,
we
aimed
to
assess
the
application
three
state-of-the-art
prompt-based
large
language
models
(LLMs),
GPT
3.5,
4,
LLaMA-2-70B,
automate
identification,
interpretation,
extraction,
structuring
relevant
information
from
textual
sources.
We
focused
on
species
distribution
two
sources:
news
outlets
papers.
assessed
LLMs
four
key
tasks:
classification
documents
with
identification
regions
where
are
recorded,
generation
geographical
coordinates
these
regions,
supply
results
in
structured
format.
4
consistently
outperformed
other
models,
demonstrating
high
capacity
interpret
extract
information,
percentage
correct
outputs
often
exceeding
90%
(average
accuracy
across
87–100%).
Its
performance
also
depended
source
type
task,
better
achieved
reports,
reports
presentation
output.
predecessor,
exhibited
reasonably
low
all
tasks
sources
81–97%),
whereas
LLaMA-2-70B
showed
worst
(37–
73%).
These
demonstrate
benefit
integrating
into
assimilation
workflows
essential
tools
efficiently
process
volumes
data.
Language: Английский
Improving Our Understanding of a Cryptic Primate, the Philippine Tarsier (Carlito syrichta), Through Social Media
International Journal of Primatology,
Journal Year:
2025,
Volume and Issue:
unknown
Published: May 8, 2025
Language: Английский
Large language models overcome the challenges of unstructured text data in ecology
Ecological Informatics,
Journal Year:
2024,
Volume and Issue:
82, P. 102742 - 102742
Published: Aug. 2, 2024
The
vast
volume
of
currently
available
unstructured
text
data,
such
as
research
papers,
news,
and
technical
report
shows
great
potential
for
ecological
research.
However,
manual
processing
data
is
labour-intensive,
posing
a
significant
challenge.
In
this
study,
we
aimed
to
assess
the
application
three
state-of-the-art
prompt-based
large
language
models
(LLMs),
GPT-3.5,
GPT-4,
LLaMA-2-70B,
automate
identification,
interpretation,
extraction,
structuring
relevant
information
from
textual
sources.
We
focused
on
species
distribution
two
sources:
news
outlets
papers.
assessed
LLMs
four
key
tasks:
classification
documents
with
identification
regions
where
are
recorded,
generation
geographical
coordinates
these
regions,
supply
results
in
structured
format.
GPT-4
consistently
outperformed
other
models,
demonstrating
high
capacity
interpret
extract
information,
percentage
correct
outputs
often
exceeding
90%
(average
accuracy
across
87–100%).
Its
performance
also
depended
source
type
task,
better
achieved
reports,
reports
presentation
output.
predecessor,
exhibited
slightly
lower
all
tasks
sources
81–97%),
whereas
LLaMA-2-70B
showed
worst
(37–73%).
These
demonstrate
benefit
integrating
into
assimilation
workflows
essential
tools
efficiently
process
volumes
data.
Language: Английский
Using social media image to identify threatened elasmobranch species caught by a small-scale fishery in a data-poor area.
Ocean & Coastal Management,
Journal Year:
2024,
Volume and Issue:
254, P. 107202 - 107202
Published: May 25, 2024
Language: Английский
Integration of ecological indicators to assess a multitemporal impact of cement industries
Environmental Science and Pollution Research,
Journal Year:
2024,
Volume and Issue:
31(35), P. 48233 - 48249
Published: July 18, 2024
The
present
study
evaluated
an
integrated
biomonitoring
approach
based
on
three
different
bioindicators:
tree
rings,
lichens,
and
beetles
in
a
complex
environment
(urban-industrial-forest).
In
Central
Italy,
four
sampling
sites
were
selected
to
assess
the
anthropogenic
impact
of
cement
plants
taking
into
account
(1)
long-term
exposure
(1988-2020)
through
analysis
trace
elements
rings
Quercus
pubescens;
(2)
medium-term
(2020-2021)
thalli
(outermost
portions)
lichen
Xanthoria
parietina;
(3)
short-term
spring
2021
bioaccumulation
evaluation
sample
vitality
transplants
Evernia
prunastri
periodic
survey
entomological
biodiversity
carried
out
during
summer
2021.
Trace
industrial
origin
found
with
levels
accumulation
between
1988
2020
maximum
2012.
Native
X.
parietina
showed
overall
low
except
for
Cr,
probably
reflecting
influence
national
lockdown
measures.
E.
weak
stress
response
urban
sites,
but
not
forest,
identified
Tl
V
as
main
contributing
atmospheric
contamination,
peaks
at
sites.
Concerning
beetles,
significantly
lower
number
species
was
Semonte
site.
Language: Английский