Recent trends in phytoplankton
populations in Lake Kinneret:
Alternate stable states hypothesis
Tamar Zohary, Kinneret Limnological Laboratory, IOLR
[email protected]
Daniel Roelke, Texas A&M University
K David Hambright, University of Oklahoma
Photo: Matthew Hipsey, May 2005
Lake Kinneret phytoplankton
“One of the best-known and best-attested
examples of year-to-year similarity in the
abundance, distribution and composition of
the phytoplankton.”
CS Reynolds, 2002
Representative P. gatunense Bloom
Typical of L Kinneret till ~ 1994
400
Biomass, g m-2
Peridinium gatunense
200
5 µm
Other Phytoplankton
0
Jan.
June
Jan.
Deviations from the stable pattern since ~1994
Biomass, g m-2
400
• No-bloom years
• Invading species
• Cyanobacteria more prominent
•200Different summer assemblage
(Zohary 2004 FWB)
0
Jan.
June
Jan.
N2 fixing cyanobacteria invaded L Kinneret
Aphanizomenon ovalisporum, 1994
Cylindrospermopsis cuspis, 2000
The summer assemblage
1969 - 1993
Small, round-shaped
Solitary & coenobia
Recent years
Filamentous, spiny,
elongated or colonieal
Recurring P. gatunense blooms
400
1969
gp-4
1970
gp-4
1971
gp-2
1972
gp-3
1973
gp-2
1974
gp-2
1975
gp-2
1976
gp-2
1977
gp-2
1978
gp-2
1979
gp-3
1980
gp-3
1981
gp-3
1982
gp-2
1983
gp-2
1984
gp-3
1985
gp-3
1986
gp-3
1987
gp-4
1988
gp-2
1989
gp-2
1990
gp-3
1991
gp-3
1992
gp-3
1993
gp-3
1994
gp-4
1995
gp-4
1996
gp-1
1997
gp-1
1998
gp-4
1999
gp-2
2000
gp-1
2001
gp-1
2002
gp-2
200
0
biomass, g m -2
200
0
200
0
200
0
200
0
Time
“Big-Bloom” Years:
1994-95, 1998
2003, 2004
“No-Bloom” Years:
1996-97, 2000-01
2005
Bloom vs. no-bloom years since 1994
What could cause the lack of
blooms on some years ?
Lake Kinneret Data Set, 34-Years (1,057)
Weekly and bi-weekly sampling
Phytoplankton
(integrated over mixed layer)
(128)
Enumeration at genus- and species-level
Zooplankton (35)
Enumeration at genus- and species-level, and life stages
Physicochemical (22)
1. Nutrients (NH4; NO3; NO2; DON; TDP; TP; SiO3; and SO4 )
2. Hydrological (inflow; flushing; loading of NH4, NO3, DON, TDP,
TP, and SO4)
3. Physical
(lake level; thermocline depth; Secchi depth; light penetration to
mixing depth ratio; temperature; and conductivity)
Simple Model Explanations
Compare early-season P. gatunense with:
1. Early-season physicochemical parameters
2. Early-season zooplankton parameters
Multivariate
Principal Component Analysis
(trends <20% of variability, contradictory trends on pcs)
Univariate
Linear and Curvilinear Regression
(poor R2 values, typically >0.10)
Proposed Complex Model Explanation
Distinct community structure can occur in the same environment
Alternate
Stable-States
Model
Figure modified from
Scheffer, 1998;
Carpenter, et al., 2001
Alternative
Equilibria in
Shallow Lakes
M. Sheffer, et al. 1993.
TREE 8: 275–279
t
nu
s
nt
rie
turbidity
Objectives and Approach
Explore whether distinct phytoplankton community states
exist in Lake Kinneret
1. Discriminant Analyses
2. Principal Component Analysis
use two-month “Running-window”
Determine whether the occurrence of these states “fit” into
an alternate states model
1. “More-Dynamic” Variable?
2. “Less-Dynamic” Variable?
Representative Discriminant Analysis (Regions X and Y)
Phytoplankton (April-May)
Bloom
years
{
gp2 - small
gp3 - medium
gp4 - big
Distinct if:
1. Six unit separation
between group
centroids
2. No overlap between
group members
Group 1 is distinct from
groups 2, 3, and 4
gp1 - No-bloom
years
Groups 2, 3, and 4 are
not distinct from each
other
Discriminant Analyses, All Years
(Regions X and Y)
Phytoplankton
April-May
gp1 distinct maximum
Jan-Feb
gp1 1st distinct
gps 2, 3, 4 - never become distinct
Phytoplankton Community Structures (Regions X and Y)
Principle component analysis
Spring
Phytoplankton
Select for
no-bloom year
Select against
no-bloom year
Chodatella sp.
Closterium acutum
Oocystis sp.
Cyanodictyon imperfectum
Microcystis aeruginosa
Microcystis wesenbergii
Carteria cordiformis
Peridiniopsis elpatiewski
Proposed Complex Model Explanation
Alternate
Stable-States
Model
No-bloom Condition
Bloom Condition
Representative Discriminant Analysis (Regions X and Y)
Zooplankton (April-May)
Bloom
years
{
gp2 - small
gp3 - medium
gp4 - big
Distinct if:
1. Six unit separation
between group
centroids
2. No overlap between
group members
Group 1 is not distinct
from groups 2, 3, and 4
gp1 - No-bloom
years
Groups 2, 3, and 4 are
not distinct from each
other
Discriminant Analyses, All Years
However …
Zooplankton
gps 1, 2, 3, and 4 never became distinct
Physicochemical
gps 1, 2, 3, and 4 never became distinct
But what if …
Zooplankton and/or Physicochemical
were a “More-Dynamic” Variable
(Regions X and Y)
Proposed Complex Model Explanation
Pre-1994
Post 1994
No-bloom Condition
Bloom Condition
Alternate
Stable-States
Model
“Lumping” data from
Regions X and Y
would mask correlation
between “Condition” and
the “More-Dynamic”
Variable
Representative Discriminant Analysis (Region Y)
Zooplankton (Jan-Feb)
Distinct if:
No-bloom
and
Bloom
years are
distinct
1. Six unit separation
between group
centroids
2. No overlap between
group members
Representative Discriminant Analysis (Region Y)
Physicochemical (Jan-Feb)
Distinct if:
No-bloom
and
Bloom
years are
not
distinct
1. Six unit separation
between group
centroids
2. No overlap between
group members
Discriminant Analyses, 1994-2001
(Region Y)
April-May
zooplankton distinct max
Jan-Feb
zooplankton distinct
Assume Causation
(Region Y)
Early-Spring
Zooplankton
Spring
Phytoplankton
Select for
no-bloom year
Ceriodaphnia rigaudi
Eudiaptomus dreischi nauplii
Aneuropsis sp.
Synchaeta pectinata
Hexarthra sp.
cyclopoid copepod nauplii
Chodatella sp.
Closterium acutum
Oocystis sp.
Cyanodictyon imperfectum
Microcystis aeruginosa
Microcystis wesenbergii
Carteria cordiformis
Select against
no-bloom year
Chydorus sphaericus
Collotheca sp.
Moina rectirostris
Peridiniopsis elpatiewski
Proposed Complex Model Explanation
Zooplankton Community Structure
Structure
2
Structure
1
Alternate
Stable-States
Model
No-bloom Condition
Bloom Condition
“Hysteresis” would
mask correlation
between “Condition”
and the “Less-Dynamic”
Variable
State of the Kinneret Fishery
Smaller bodied zooplanktivores
exert greater feeding losses to
zooplankton
Zooplankton communities might
respond by getting smaller
State of the Kinneret Zooplankton
ug indivual (wet weight)
Ceriodaphnia sp.
Members of the zooplankton
community do get smaller!
Smaller zooplankton likely
exert greater grazing losses
to phytoplankton, which might
select for more quickly
growing phytoplankton taxa
Proposed Complex Model Explanation
Zooplankton Community Structure
Structure
Two
Structure
One
Alternate
Stable-States
Model
No-bloom Condition
Bloom Condition
Large
Small
Fish and Zooplankton Body Size
Phytoplankton Phase Space - Hypothesis
When fish and
zooplankton body
size are “large”,
only these trajectories
are possible
When fish and
zooplankton body
size are “small”,
both trajectory paths
are possible Zooplankton community
structure is the trigger