Transcript Datasets introduced

Datasets introduced
Traits DGS
2/14/2008
Cornwell data set: Woody plant species of coastal CA,
across gradients in elevation and insolation
Objective: test for trait-based patterns in community assembly
Site: Jasper Ridge Biological Preserve, Stanford Univ. (37.4°N, 122.25°W)
Sampling: 44 20 x 20 m plots randomly located across 5 vegetation types; relative
abundance of all woody plants sampled
Species: 54 species total, 3-18 spp. per plot
Environment: elevation, relative insolation (based on topography), soil moisture,
and soil nutrients measured in each plot
Traits: 14 leaf, stem and root traits
measured for (almost) all spp.; SLA and
leaf size measured in situ in each plot
Quic kTime™ and a
TIFF (LZW) decompress or
are needed to see this pic ture.
Will Cornwell, post-doc, Univ. British
Columbia, [email protected]
Cornwell data set: Traits and results
11 continuous, above-ground
traits
Leaf size
Specific leaf area (SLA)
Nitrogen/leaf mass (Nmass)
Nitrogen/leaf area (Narea)
Leaf area:sapwood area ratio
(LA:SA)
Wood density
Vessel size
Vessel density
Vessel lumen fraction
Plant height
Seed mass
3 root traits:
N-fixer
AM myc. association
EM/arbutoid myc. association
• On average, individual plots have a smaller multi-dimensional range
for leaf size, SLA, wood density and seed size, evidence of
environmental filtering (Cornwell and Ackerly 2006, Ecology)
• Method devised for partitioning species trait values into two
components, one associated with distribution along gradients, and the
other reflecting within community variation (species trait value relative
to co-occurring species) (Ackerly and Cornwell 2007, Ecology Letters)
• SLA values are more evenly spaced within communities than
expected under a null model, evidence of limiting similarity in the
assembly of communities (Cornwell and Ackerly 2008, Ecol. Monogr.)
• Mean trait values are correlated with underlying moisture gradient;
wood density higher in dry sites, SLA higher in moist sites. LA:SA and
wood density had higher variance in moist sites, and Narea higher
variance in dry sites; evidence of shift in traits associated with niche
partitioning in different environments (Cornwell and Ackerly 2008,
Ecol. Monogr.)
Ecological Flora of California
Prototype plant trait database
Open access, fully searchable and downloadable
Currently has 3278 records for seed size of 2002 species,
and 2370 records for 458 taxa for:
Blade area (c)
Specific leaf area
Blade length (c)
Blade width (c)
Leaf thickness
Leaf type
Wood density
Leaf Nmass
Maximum height (c)
Leaf area to sapwood area
Leaf Narea
Annual seed production per plant
Let’s add much more data!
mysql prototype: http://bscit.berkeley.edu/efc/
Kunza vs Konza
• Please note the difference between the
Kunza dataset (Amy Kunza’s work on salt
marshes) and the Konza LTER site.
Kunza and Pennings. Plant composition of GA and TX salt marshes.
• 49 (TX) and 59 (GA) sites
– Sites scored for landscape position
(mainland, hammock, barrier island)
and upland influence (bay to
peninsula on 5 point scale)
Upper edge of the
marsh
• Transects from water to marsh edge
with sets of plots at 1m intervals
Transects average
ca. 100m
– Species presence and percent cover (only
0.5 x 0.5 m plots)
• Each set of plots consists of 9 nested
sub-plots (that can be used to
generate a species-area curve)
1.
2.
3.
4.
5.
6.
7.
8.
9.
0.1 x 0.1m
0.1 x 0.25m
0.25 x 0.25m
0.25 x 0.5m
0.5 x 0.5m- percent cover
0.5 x 1.0m
1.0 x 1.0m
1.0 x 2.5m
1.0m
1.0 x 5.0m
Water’s edge (as close as possible)
5.0m
GA vs. TX
• Limited species pool (43)
– Similar regional richness (GA: 27; TX: 32)
– Overlap in species (16, including most abundant species, were
shared)
• General climate
– Annual temperature and climate similar (w/in 5o & 5”)
– Same NEON climate domain
• Same salt marsh habitat sampled in both states
– Share common stress gradients: waterlogging, salinity, anoxic
soils
– But predictability and periodicity of tides differ
• GA Plant zonation stricter, plots and sites less diverse
– GA: 1.67 species/plot, 5.83/site
– TX: 2.85 species/plot, 9.35/site
• Dataset was used for Kunza MS thesis; now freely
available
– (Kunza and Pennings, submitted to Estuaries and Coasts)
ORIGINAL QUESTIONS TO BE ADDRESSED BY THE IRRIGATION STUDY:
Can site productivity be maintained under an annual fire regime with net
primary production maximized each year? Or will soil N limitations become
more severe over time as predicted by Century model output (Ojima et al.,
1990)?
Will reproductive effort of the dominant grasses be enhanced by long-term
irrigation? Or will allocation to reproduction decline as belowground
resources become limited?
Will significant plant species composition shifts occur? More specifically, is
annual fire sufficient to keep woody species from increasing along these
wet transects? Will increased competition for N and light reduce the
abundance of forbs?
How will soil processes such as decomposition and N mineralization be
affected by irrigation? Will readily mineralizable C and N pools decrease
(Ojima et al., 1990)? How will the temporal dynamics of microbial biomass
be affected?
See Knapp et al. 2001 Ecosystems,and on-line methods document.
VARIABLES MEASURED
1) Aboveground biomass
2) Plant reproductive effort
3) Xylem pressure potential in Andropogon gerardii
4) Plant species composition
5) Soil moisture and chemistry
http://www.konza.ksu.edu/
Dataset WAT01
Irrigation transects
Irrigated
10 m2
Control
10 m2
Vegetation was
sampled from 19912006 in 31
permanently located
10-m2 quadrats along
two irrigated and two
control transects
NPP is estimated
annually by clipping
20 x 50 cm quadrats
adjacent to each
species composition
plot
10-m2 species
comp quadrat
1
Lowland
16
1 (quadrat number)
16
20x50 cm
NPP clip
quadrat
Control
Irrigated
5
Irrigated
5
Control
20
20
Steep rocky slope
10
10
27
27
15
15
Upland
31
31
10
Upland irrigated
9
Upland control
Lowland irrigated
8
antilog(H')
Precipitation
effects on
species
diversity
Lowland control
7
6
5
4
3
1990
1992
1994
1996
1998
2000
2002
Year
Irrigation had no significant effect on
species diversity in upland or lowland prairie
20
Upland control:
r=0.55, P=0.10
Total richness
18
Precipitation
effects on
species
richness
over time
Upland irrigated:
r=0.62, P=0.06
16
14
Lowland irrigated:
r=0.77, P=0.009
Lowland control:
r=0.74, P=0.02
12
10
200
400
600
800
1000
1200
1400
Total precipitation
• Species richness increases with increasing total
precipitation
• The response is stronger in lowland compared to upland
soils
Konza Prairie Long-term Irrigation Transect Study
• Treatments initiated in 1991
• Supplemental water added during the growing season to
replicate 140 m transects (paired with control transects)
• Designed to meet plant water demand and minimize
intra-annual variability in soil water deficits
Summary of the first eight years…
• Water availability limited
ANPP 6 out of 8 years
900
Irrigated
2
Total ANPP (g/m )
Control
800
• Irrigation increased ANPP
by ~25% (physiological
response)
700
*
600
*
500
400
*
*
*
*
300
1200
800
600
800
400
200
0
r2 = 0.81
1992
2
ANPP (g/m )
700
1994
1996
1998
Year
600
500
400
300
600
800
1000
1200
Annual ppt + irrigation (mm)
1400
• Good fit between ANPP
and ppt amount and when
variability is removed and
range extended
Knapp et al. 2001 Ecosystems
PPT (mm)
1000
But then we continued the experiment…
• Driven by responses of
grasses
900
Control
800
Irrigated
2
Total ANPP (g/m )
• Mean increase in ANPP for
the next five years was
70%
700
*
*
600
*
500
400
*
*
*
*
*
*
*
*
300
1200
800
900
ANPP (g/m2)
800
600
1999-2003
irrigated
2
400
200
r =0.52
PPT (mm)
1000
0
1992
700
1994
1996
1998
2000
2002
Year
600
1991-1998
irrigated
500
1991-2003
control
400
300
600
800
1000
1200
Annual ppt + irrigation (mm)
1400
Changing relationship between
ANPP and precipitation after
long-term increase?
400
What would change the
relationship between ANPP
and precipitation?
= 1999-2003
300
r2 = 0.66
200
= 1991-1998
100
0
0
100
200
300
400
500
Irrigation water added (mm)
Andropogon gerardii
control
irrigated
80
• May be related to species
changes (increased cover
of Panicum virgatum) –
Community response
Average cover (%)
2
Stimulation in ANPP (g/m )
500
60
40
Panicum virgatum
control
irrigated
20
0
1992
1994
1996
1998
Year
2000
2002
EXPERIMENTAL DESIGN
KONZA PRAIRE RESEARCH 1999
WATER SUPPLEMENTATION
4
sprayer
UPLAND TRANSECT
85% water
85%
3
2m
0 gN
2
50% water
0 gN
0 gN
Hundreds
Depth of Water Added, mm
50%
2.5 gN
2.5 gN
1
0%
0
-20
0% water
2m
NORTH
5 gN
-10
0
10
Distance SOUTH of Sprinkler Line, m
Fig. 6. 0 on the x-axis refers to the position of the sprinkler line.
Bars indicate water supplementation used in this experiment
(in mm water) and as % of control.
2.5 gN
5 gN
5 gN
10 gN
10 gN
20
10 gN
BLOCK REPEATED 6 TIMES IN UPLAND AND LOWLAND,
ACTUAL LOCATION OF N ADDITIONS ARE RANDOMIZED.
Treatment effects on dominant grasses
100
100
Average cover
Big bluestem
Switchgrass
80
80
a
b
60
40
60
40
30
30
c
20
10
a
b
b
c
b
20
10
a
0
0
0.0
2.5
5.0
10.0
Nitrogen (g m-2)
Upland
Lowland
Site
•Resource addition alters species interactions
•Strength of response is context dependent
Tiffany Troxler
Freshwater marsh
LOX
Live biomass (g/0.25m2)
Dead biomass (g/0.25m2)
Number of species/m2
2A
Total stems/m2
Species (stems/m2)
Transects (12-16 km)
Anchored at
Canal (phosphorus)
inflow points at
5 sites
3A
SRS
1989
1999
current data?
TS
1999 tissue CNP of
3 species
3 sites in SRS, flume experiment
P added at three levels over 5+ years
Species plot data
(composition/height) along flume
“transects”-from inflow point; also
water quality, soil nutrients, other
At least one site in TS, plot data from
canal outflow point – likely since
1999, water quality, soil nutrients,
tissue nutrients of Cladium at least
LTER data –
•Cladium, SRS 1, 2, 3, bimonthly, density, biomass, tissue nutrients; 2000-present + soil
nutrients + hydrology, salinity
Mangrove forest
LTER data –
•Mangrove composition, density, DBH, height
(intermittenly since 1996), tissue CNP, soil CNP,
SRS 4, 5, 6; SRS 4 and 6 annually since 1991; soil
nutrients, hydrology, salinity
Freshwater wetland forest and marsh
Peat development gradient – Space for time substitution?
Raphia palm forest along coast
Interior bog plain
~4000 yrs
Mixed hardwood
SRS also, I believe paleo data exists
~800 yrs
Soil nutrients, tissue C & nutrients for dominant species (canopy and understory),
species composition, enzyme activity, CO2 flux, DBH for canopy species in 5 0.25
ha plots along gradient
I. Variation in soil nutrients with distance along peat
development gradient
1500
r2 = 0.971
Soil Phosphorus
y = -362.48x + 1284.14
F = 200.94; p < 0.0001
ug TP/g soil
1250
1000
750
500
250
0
0.0 0.3 0.6 0.9 1.2 1.5 1.8 2.1 2.4 2.7 3.0
Coast
distance (km)
Inland
Synthesis of N fertilization experiments in North America
ARC -Arctic
Tundra
CDR -Oak
Savanna
JRG - Annual
Grassland
CAR -Pacific
Saltmarsh
KBS -Abandoned
Agricultural Field
SGS-Shortgrass
Prairie
NWT -Alpine
Tundra
SEV -Desert
grassland
KNZ -Tallgrass
Prairie
GCE -Atlantic
Salt Marsh
Synthesis of N fertilization
experiments in North America
-Multiple experiments at each
site for a total of 37 experiments
ARC
-Encompassed the abundance
responses of 640+ species
CDR
-Each species was classified by
traits:
KBS
SGS
JRG
NWT
CAR
SEV
-Origin (native, non-native)
KNZ
-Life form (Grass, forb, shrub)
GCE
-Photosynthetic pathway (C3, C4)
-Potential for N fixation
-Height class in canopy
-Life history (annual, perennial)
Cleland et al. (2008) Ecology - data publication
-Clonality (tussock, rhizome)
FertSyn dataset
• Metadata and data available at:
http://traitsdgs.nceas.ucsb.edu/workspace/data/datasets/database-of-n-fertilizationexperiments/
Cedar Creek - plant traits and community
composition along an old field
chronosequence
• Collected by Claire Claire
Jouseau, Kally Worm,
Shahid Naeem, Jean Knops
and Dan Bunker
• Community composition
– 11 old fields - 10-80
years old
– 4 100m transects per
field
– 10’s of 1m neighborhood
plots per transect with %
cover of all species
recorded
– 5 years of data
• Plant traits for 50 most
common species, n=6
– Leaf %C, %N, LMA, area
– Height
– Root:stem:leaf:reproductive
biomass
– Longevity
– Functional Group (C3, C4,
forb, legume)
– Nativity
– Seed mass
– Light response curves for
25 most common
Additional resources
• Kew Gardens seed
database
• Glopnet
• VegBank
• USDA plants
• USFS Forest
Inventory
Alpine Trait Dataset: K. Suding ([email protected]) and I. Ashton
• How do changes in nutrient availability affect N cycling and
community composition? An experiment manipulation of nitrogen
availability (add, ambient, reduce) since 2000. Also in a factorial
design with species removals, but we did not measure traits in the
removal plots.
2) Can species traits predict abundance or response to nutrient
treatments? In 2006, we measured SLA and C:N.
3) What level of trait variation is important (within species, within
species at each plot, within individuals at each plot)? We measured
these traits for three individuals of every species present in every
treatment plot (660 individuals were measured, 31 species).
Experimental Design
DATA SET
• Seven replicate sites, 1 m2 plots
• Nutrient additions since 2001
•Nitrogen fertilization
(increased N)
•Carbon addition (reduce N)
• Annual measure of species
composition & resin available N
• In 2006, collected trait data from
all species (Control, +C, +N) at the
plot level
• 3 individuals/species/plot for SLA
• 3 individuals/species/plot pooled
for C:N data
•For a total of trait measures on
660 individuals and 31 species
Results: How do changes in nutrient availability affect
community composition?
BACKGROUND RESULTS:
Composition and N cycling effects are in press in Ecology. Briefly, the N
manipulations were effective in increasing and reducing N supply. N addition
caused a large decline in one of the dominance species (Acomastylis rossi) and a
subsequent increase in some graminoid species. No large changes in overall
diversity, but some subordinate species did respond.
Availability/Use
• You are welcome to use the trait dataset anyway that a DGS group wants.
• We can provide species composition (2002-06), nitrogen, and other related
data, and are interested to help.
• Isabel Ashton is a postdoc at UCI and collected these data
([email protected]). Any problems with the experimental design are my fault
([email protected]).