Transcript PowerPoint - Carolina Vegetation Survey

Components of plant species diversity in
the New Zealand forest
Jake Overton
Landcare Research
Hamilton
Acknowledgements
NVS data contributors and curators
Simon Ferrier and Glenn Manion for development
of GDM and collaboration on modelling
General Question
Investigate components of richness
• Alpha diversity
• Beta diversity
• Gamma diversity
How do these compare between groups?
Approach: Use a new modelling technique, Generalised
Dissimilarity Modelling (GDM) to estimate components
of diversity
Components of diversity (sensu Cody
1986)
Alpha diversity = local richness
Beta diversity = turnover in species due to habitat or
environment
Gamma diversity = turnover in species due to
geographic distance or barriers
All three components contribute to regional richness
Biotic
Data
NVS recce
(= recon)
plots
Almost 20000
plots
1220 species
Presenceabsence of all
vascular plant
species in
each plot
Plots approx
20x20 m
(sometimes
unbounded)
Environmental variables (spatial)
Variable abbrev.
Description
Geographic position
Geographic position of plot
MAT
Mean Annual Temperature
Tseas
A measure of cold stress, relative to mean
annual temperature
MAS
Mean Annual Solar Radiation
Deficit
Vapor Pressure deficit
VPD
Vapor Pressure deficit
Calcium
Soil Calcium
Age
Soil age
N
Total Soil Nitrogen
AcidP
Available P
Drain
Soil Drainage
Psize
Soil Particle Size
Indur
Soil Induration
Slope
Topographic slope
Discoast
Distance to coastline
Notho
Nothofagus abundance from Leathwick
What is Generalised Dissimilarity Modelling?
alpha diversity (local richness)
beta diversity
gamma diversity
‘dissimilarity’
‘turnover’
‘complementarity’
Modelling of richness:
richness = f (rainfall, temperature, veg type …)
can be supplemented by modelling of compositional dissimilarity between locations:
dissimilarity = f (
(rainfall, temperature, veg type …), geographical separation)
Generalised dissimilarity modelling (GDM)
 ln 1  dij      f k xki   f k xkj 
p
k 1
Compositional
dissimilarity between
pairs of survey
sites
Environmental
& geographical
separation
Biotic Information
Differing units and
importance
Environmental and Geog Space
Ecological Space
Same units
scaled by importance
All species model
All species validation
Results 1
 ln 1  dij      f k xki   f k xkj 
p
k 1
Unexplained component =
1 – proportion deviance explained
Alpha diversity component =
Proportion accounted for by local richness =
Mean plot richness/ Total species pool
Gamma Diversity component =
Proportion deviance explained by
geography
Beta Diversity component =
Deviance explained by environment
Total species pool
1020 species
All species
All plant species
Snails
Group
# spp
local rich
Ferns
147
6.8
Lianes-epiphytesparasites
49
1.9
Trees
107
8.9
Shrubs
255
4.3
Dicot herbs
360
2.4
Monocot herbs
282
1.3
All Species
1020
25.9
Results 1
All
Shrubs
Ferns
Trees
Monocot herbs
Dicot Herbs
All species
Ferns
Trees
Shrubs
Monocot herbs
Dicot Herbs
Predicted distributions
of species
Constrained
environmental
classification
Biological
survey
data
Generalised
dissimilarity
modelling
Visualisation of spatial
pattern in community
composition
Conservation
assessment
Environmental
predictors
Climate-change
impact assessment
Survey gap
analysis
Ferrier, S. et al (in press) Using generalised dissimilarity modelling to analyse and predict patterns of beta-diversity
in regional biodiversity assessment. Diversity & Distributions
Ferrier, S. et al (2004) Mapping more of terrestrial biodiversity for global conservation assessment.
BioScience 54: 1101-1109
Conclusions
GDM is an exciting new tool for biodiversity analyses
Its main application is for biodiversity modelling and
planning, but it has promise for untangling components
of diversity
Plant species show relatively strong environmental
influence and some geographic influence on turnover
Groups differ in the explained turnover, and in relative
importance of different variables.
test
Sparse sampling relative to grain of
compositional turnover - huge number of
species, each with very few (or no) records
Geographical space (gamma diversity)
Geographical space (gamma diversity)
Dense sampling relative to grain of
compositional turnover - relatively few
species, each with many records
Environmental space (beta diversity)
Environmental space (beta diversity)
An example from the arid rangelands of central Australia
All species
All species
400
600
800
Biological response
1000
1200
1400
Elev ationd
0.4
0.4 0.6
0.6 0.8
0.8 1.0
1.0
0.0
0.0
k 1
0.2
0.2
 ln 1  dij      f k xki   f k xkj 
p
f (Tc10d)
Predicted
Response
Predicted
Response
Bray-Curtis compositional dissimilarity
between all pairs of 248 field survey sites
(based on perennial woody plant species)
Environmental predictors
262
500
1000 2641500
2000266 2500
268
3000
0.0
0.2
0.4
0.6
0.8
1.0
Eacasp26d
Tc10d
f (Wetness)
Predicted
Response
•Radiometrics – Total Count
•Landsat TM – Band 2
•Radiation of Warmest Quarter
•Topographic Wetness Index
•Precipitation of Driest Period
•Isothermality
•Minimum Temperature of Coldest Period
•Elevation Diversity for 300m radius
•Landsat TM – PD54 vegetation index
•Mean Temperature of Wettest Quarter
•Radiometrics – Uranium
0
2000
4000
6000
8000
Wetness
10000
12000
What is Generalised Dissimilarity Modelling?
Models species turnover (dissimilarity) between locations as
a function of geography and environment
Uses matrix regression, using GLMs.
Developed by Simon Ferrier, (Department of Environment
and Conservation, Armidale New South Wales, Australia)
Programmed by Glenn Manion, DEC, Armidale.
All species
Ferns
Trees
Shrubs
Monocot herbs
Dicot Herbs
Ferns
Ferns
L-E-P
test
Monocot herbs
Monocot herbs
Shrubs
Shrubs
Trees
test
Dicot Herbs
Ferns
L-E-P
test
Monocot herbs
Shrubs
Trees
Trees