Transcript Habitat loss and fragmentation II

Habitat loss and
fragmentation II
Bio 415/615
Questions
1. What changes in patch structure occur with
habitat fragmentation?
2. How is fragmentation measured?
3. Are Eastern U.S. forests more or less
fragmented than they were 150 years ago?
4. What does an incidence function measure?
5. What is remote sensing, and how can it
improve deforestation rate measurements?
Example of
Habitat loss and
Fragmentation in a
Wisconsin Forest
Brazilian Atlantic
Rainforest
Cote d’Ivoire Forest
Norway
1900
1940
Wilderness
>5 km from human centers
1998
Fragmentation
•
•
•
•
Total area decreases
Size of patches decreases
Number of patches increases
Isolation increases, connectedness
decreases
• Edge length increases
• But changes in habitat arrangement
rarely occurs without habitat LOSS
How is fragmentation measured?
•
•
•
•
•
•
•
Number of patches
Mean patch size
Probability of adjacency
Average proximity
Connectivity
Shape complexity
Edge length
Understanding fragmentation
• Relevance: some habitats and populations are
naturally patchy
• Age and dynamics over time
• Matrix properties
• Ecological contrast (human patchiness vs.
natural heterogeneity)
• Goldilocks principle: specialists and generalists
• Local/Regional scaling issues (rescue effect,
dispersal limitation)
• Edge effect
Eastern U.S.: forests to fields and back again
Current diversity bears the imprint of
past fragmentation (Vellend 2003)
The extent of past forest loss influences the present richness of
forests, even after widespread reforestation… the processes of
re-equilibration can take centuries!
Tropical deforestation
• Greatest threat to global biodiversity?
• How do we measure deforestation rate?
– Reporting by gov’t organizations and NGOs
(UN Food & Agriculture Organization, State of the
World’s Forests)
Biases?
– Remote sensing
FAO data: how to interpret?
• Grainger 2008 PNAS on FAO Forest Resources Assessments
(FRAs): “it is difficult to demonstrate [tropical forest decline]
convincingly using… available data”
• Large variance in estimates of tropical forest area; FAO tends
to be high
• From estimates below, do we conclude that tropical forests are
actually increasing? NO!
• Errors include 1) subjectivity in ‘expert assessments’ 2)
reporting is done in different years among countries, so data
must be corrected to a standard year 3) forest classes change
over time and must be lumped, separated, etc (open forest,
savanna, closed canopy, etc)
Remote sensing
This map shows percent forest cover in eastern
Democratic Republic of the Congo. Red areas,
expanding outward from the town of Mambasa, show
deforestation between 1990 and 2000. (Map by
Robert Simmon, based on data from the Decadal
Forest Change Mapping Project.)
NASA Earth Observatory
Some areas of Rondonia, Brazil, have been almost completely
deforested in just 6 years. This pair of images uses a scale, or
index, of vegetation to compare forest area in 2000 to 2006 at the
full resolution (15 meters per pixel) of the Advanced Spaceborne
Thermal Emission and Reflection Radiometer (ASTER)
instrument. Cleared areas (tan) spread from roads cut through
the forest (green), a pattern of deforestation typical in Rondonia.
(Maps by Robert Simmon, based on ASTER data.)
NASA Earth Observatory
NDVI = normalized difference
vegetation index
NDVI = near-infrared – red
near-infrared + red
• Chlorophyll absorbs (visible) RED light
(and blue light) and scatters nearINFRARED light (>800 nm).
• So plants look bright in infrared, but
dark in red. The difference between
the intensity of these spectral
measurements indicates how much
chlorophyll is on the ground, and is thus
a gross measure of plant density (e.g.,
leaf area index or LAI).
Estimating tropical deforestation rates:
Achard et al. 2002 Science
• Closed evergreen forests in humid tropics
• Change in cover of several forest type categories
between 1990 and 1997 via remote sensing
• About 6 million ha lost each year globally; another .7
million ha degraded
• BUT this was 23% LESS than the FAO estimate!
loss
gain
Net
change
Projecting future deforestation rates:
Soares-Filho et al. 2006 Nature
• Trends in agricultural
expansion in the Amazon basin
• Based on current growth
trends in cattle ranching and
soy production
• 40% loss by 2050
• Expansion and enforcement of
current protected areas could
reduce loss by 1/3
Regions of loss
by 2050
Rate of loss by
ecoregion
What happens as a result of
fragmentation?
•
•
•
•
•
Island biogeography:
Genetics:
Populations:
Metapopulations:
Communities:
What happens as a result of fragmentation?
• Island biogeography: extinction increases,
immigration decreases (paleorefuges, landbridge
islands), extinction debt is paid, nestedness increases
• Genetics: effective population size decreases, drift
increases, gene flow decreases
• Populations: population size decreases, extinction risk
increases
• Metapopulations: immigration decreases, patch
occupancy falls, extinction risk increases
• Communities: species interactions change, integrity
(area larger enough for key processes like top
predators and fire) decreases
What happens to particular species?
• Habitat preferences?
• Random vs. deterministic changes?
• Incidence functions, thresholds?
– Area-sensitivity
– Isolation-sensitivity
Effects of
Megareserves for large predators
patch size loss
>106 ha
100 km linear extent
Incidence Functions
Crooks 2002
Con.Bio. 16: 488-502.
Add isolation into incidence function
Crooks 2002
Fragmentation’s other side:
Edges, Edge Effect, Edge Influence
Species interactions and edge:
predation
Avian brood parasitism
Brownheaded
cowbird
Conservation response to
fragmentation
• Area
• Buffer…or just
edge effect?
• Corridors
– Use?
– Disease and
invaders?
Edge influence on forest structure and composition in
forested landscapes.
Harper et al. 2005. Con. Bio. 19:768-782.
Harper et al. 2005
Con. Bio. 19:768-782.
• Getting more
sophisticated about
edges: sun angle and
direction
Debinski & Holt 2000
Con. Bio. 14:342-355.
Debinski & Holt 2000
Con. Bio. 14:342-355.
• 20 studies, worldwide
– 6 in forests
– 14 in grasslands, old fields
• 6 Hypotheses
–
–
–
–
–
–
Spp richness increases with area
Spp abundance, density increases with area
Inter-spp interactions modified
Edge effect influences ecosystem services
Corridors enhance movement
Connectivity increases richness
Debinski & Holt 2000
Con. Bio. 14:342-355.
• Little consistency
– Edge effects, competitive release, spatial scale
• Transient effects; some trends evident only in
the long-term studies, >=14 yrs; Relaxation times
long
• Goldilocks problem: species and group
differences
– Scale differences
– Arthropod experiments best fit richness & area
– Highly mobile (birds, mammals), early successional
plants, long-lived spp, and generalist predators did not
respond as expected
– Movement, richness correlated with corridors,
connectivity
Ecosystem services!
Lyme Disease
‘natural’ experiment in
Venezuela; predators of
vertebrates are absent due to
fragments being too small
Results
• All islands lost jaguar, puma, harpy eagle within 4
years
• Smallest islands, <1 ha, lost 75% of their species
• Abundance increases for spp able to persist on
small islands: 2x (birds), 30x (howler monkeys),
100x (leaf cutter ants)
• Trophic effects as in HSS (the world is green):
herbivores up on small islands, but some down on
medium islands where armadillos occur and escape
their own predators
• Chance effects: the agoutis trapped on one island
converted plant cover to spp whose seeds agoutis
like to bury!
Breeding Bird Survey Data
Donovan & Flather 2002
Ecol. App. 12:364-374.
Donovan & Flather 2002
Ecol. App. 12:364-374.
Donovan & Flather 2002
Ecol. App. 12:364-374.
Laurence et al. 2001
Conserv. Biol. 16:605-618
dasdfa
(Numbers are keyed
to citations)