Mixed layer depth variability and phytoplankton phenology in the

Report
Mixed layer depth variability and phytoplankton
phenology in the Mediterranean Sea
H. Lavigne1, F. D’Ortenzio1, M. Ribera d’Alcalà2, H. Claustre1
1. Laboratoire d’Océanographie de Villefranche, France
2. Stazione Zoologica A. Dohrn, Naples, Italy
45th Liège Colloquium – May 17th 2013
INTRODUCTION
DATA & METHODS
RESULTS
DISCUSSION
CONCLUSION
• It is now well recognized that the phytoplankton phenology is
impacted by mixed layer depth (MLD) variability (blooms events are
good examples).
• However, it is still challenging to observe and characterize the
impact of MLD on phytoplankton (MLD and phytoplankton biomass
change rapidly, low availability of the phytoplankton biomass data).
• Merging in situ MLD data and ocean color chlorophyll-a
concentration ([Chl]SAT) data represents a way to explore interactions
between MLD annual cycle and phytoplankton phenology.
INTRODUCTION
DATA & METHODS
RESULTS
DISCUSSION
CONCLUSION
Present analysis was performed on the Mediterranean Sea because
 Data availability (weak cloud coverage for [Chl]SAT, and
important CTD sampling).
 contrasting biogeochemical regimes co-exist over the basin.
Present analysis was based on :
 The generation of concomitant MLD and [Chl]SAT annual
cycles.
 Spatial averages computed in established bioregions.
 Description of MLD and [Chl]SAT cycles based on a new set of
metrics.
INTRODUCTION
DATA & METHODS
RESULTS
DISCUSSION
CONCLUSION
Data
Satellite surface chlorophyll-a
concentration ([Chl]SAT)
Mixed Layer Depth (MLD) calculated from
in situ CTD measurements.
• SeaWiFS (1998 – July 2007) et MODISAqua (July 2007 - 2010)
• Historical database (D’Ortenzio et al. 2005
updated with Coriolis).
• Level 3, 8-day , 9km
• 72186 profiles of temperature and salinity
• Standard NASA algorithm (O’Reilly et al.,
2007)
• Computation of MLD (criteria in density
difference 0.03 kg m-3)
The data density is
not sufficient to
work with a regular
mesh grid.
A bioregionalization
was used instead.
INTRODUCTION
DATA & METHODS
RESULTS
DISCUSSION
CONCLUSION
The geographical framework - Starting point the bioregionalization of the
Mediterranea Sea proposed by D’Ortenzio and Ribera d’Alcalà (2009).
Result from a k-means
cluster analysis based on
the seasonal cycle of
SeaWiFS chlorophyll- a
concentration
Source: D’Ortenzio et Ribera D’alcalà (2009)
3 mains kinds of dynamics appeared
bloom
intermittent
no bloom
INTRODUCTION
DATA & METHODS
RESULTS
DISCUSSION
CONCLUSION
Data processing
Med NW - Bloom
bioregion
Ionian - No bloom
bioregion
MLD and [Chl]SAT
observations
temporal
spatial
Generation of concomitant MLD and [Chl]SAT annual cycles at 8-day resolution
Climatological scale
All MLD and [Chl]SAT
observations are mixed to produce
a climatological cycle.
Interannual scale
Data are averaged for each year
separately.
INTRODUCTION
RESULTS
DATA & METHODS
CONCLUSION
DISCUSSION
Metrics to describe annual MLD and [Chl]SAT cycles
CHL-MAX
MLD-MAX
[Chl]SAT
MLD
July year n
ΔINIT
ΔMAX
June
year n+1
CHL-MAX: annual maximum of [Chl]SAT
MLD-MAX: annual maximum of MLD
ΔINIT: Time lag between the initiation of mixing and the initiation of [Chl]SAT
increase (determination of the initiation date: annual median + 5%; Siegel et
al., 2002).
ΔMAX: Time lag between the date of MLD maxima and the date of [Chl]SAT
maxima.
INTRODUCTION
DATA & METHODS
RESULTS
DISCUSSION
CONCLUSION
The climatological scale
Med NW Bloom
Ionian - No
Bloom
MLD-MAX
CHL-MAX
ΔINIT
ΔMAX
185 m
0.99
mg m-3
48 days
48 days
MLD-MAX
CHL-MAX
ΔINIT
ΔMAX
90 m
0.22
mg m-3
16 days
8 days
INTRODUCTION
DATA & METHODS
RESULTS
DISCUSSION
CONCLUSION
The interannual scale : the analysis of annual cycles
Med NW - Bloom
• The shape of MLD and [Chl]SAT cycles vary from year to year.
• For 4 cycles out of 5, the succession MLD deepening followed by [Chl]SAT increase
and decay is repeated.
• Cycle 2006/2007 is anomalous.
Ionian – No Bloom
• The shape of MLD and [Chl]SAT cycles are fairly similar to the climatology.
• The absolute values, especially for MLD and [Chl]SAT peaks, are variable.
INTRODUCTION
DATA & METHODS
RESULTS
DISCUSSION
CONCLUSION
The interannual scale: the analysis of metrics
bioregion
MLD-MAX
CHL-MAX
ΔINIT
ΔMAX
Med NW - Bloom
368m
[119 – 524]
1.44 mg m-3
[0.84 – 1.72]
14 days
[0 – 40]
31 days
[(-16) – (+72)]
Ionian - No Bloom
108m
[76 – 158]
0.25 mg m-3
[0.19 – 0.30]
17 days
[(-8) – (+32)]
11 days
[(-56) – (+88)]
• The metric MLD-MAX is highly variable, by comparison to CHL-MAX.
• MLD-MAX and CHL-MAX are both higher in the “Bloom” bioregion than in the
“No Bloom” bioregion.
• ΔINIT is relatively stable and similar for the “Bloom” and “No Bloom” bioregions.
• ΔMAX is more variable, especially for the “No Bloom” bioregion.
• ΔMAX is higher for the “Bloom” than for the “No Bloom” bioregion.
INTRODUCTION
DATA & METHODS
RESULTS
DISCUSSION
CONCLUSION
Summary
 Metrics are a powerful tool to identify patterns in the MLD and [Chl]SAT
cycles.
 These patterns are relatively consistent between interannual and
climatological analyses.
How we can explain the ΔMAX
 Metrics analysis confirmed thatdifference
in the Mediterranean
stronger
and theSea
ΔINIT
ofbiomass
30
accumulation matches with areas where winter days?
MLDs are the deepest.
 Metrics analysis revealed temporal differences between main MLD and
[Chl]SAT events (measured with ΔINIT and ΔMAX).
c
ΔINIT
ΔMAX
BLOOM
~30 days
~30 days
NO BLOOM
~30 days
~0 days
INTRODUCTION
DATA & METHODS
RESULTS
DISCUSSION
CONCLUSION
Do light and nutrient availability can explain the ΔINIT and ΔMAX
values in the « Bloom » and « No Bloom » bioregions?
PrNUT
Probability that the MLD is deeper
than the nitracline depth.
Empirical estimation (MLD and
nitracline datasets are confronted)
Nitracline data set:
Nitracline = isoline 1µM
PrLIGHT
Probability that the MLD is above
the critical depth (Dcr, Sverdrup
1953).
Empirical estimation (MLD
dataset is compared to a
climatological estimation of the
critical depth, calculation method
Siegel et al. 2002)
1.3 mol photon m-2 d-1
Calculated from a dataset of 5318
nitrates profiles (MEDAR, SESAME
projects).
SeaWiFS climatology
INTRODUCTION
DATA & METHODS
RESULTS
DISCUSSION
CONCLUSION
Do light and nutrient availability can explain the ΔINIT and ΔMAX
values in the « Bloom » and « No Bloom » bioregions?
Ionian – No Bloom
Med NW - Bloom
Ionian – No Bloom
PrLIGHT
PrNUT
Med NW - Bloom
INTRODUCTION
DATA & METHODS
ΔINIT
♦ BLOOM : ~ 30 days
♦ NO BLOOM : ~ 30 days
Hypothesis:
[Chl]SAT increase only
when the MLD is below
the nitracline (November)
RESULTS
DISCUSSION
CONCLUSION
ΔMAX
♦ BLOOM: ~ 30 days
Hypothesis:
Episodically, a deficit of
light could limit the
growth during winter.
♦ NO BLOOM : ~ 0 days
Hypothesis:
Light is always available
and irregular nutrients
supplies by mixing sustain
phytoplankton growth
during winter.
INTRODUCTION
DATA & METHODS
RESULTS
DISCUSSION
CONCLUSION
Conclusions and Perspectives
• Metrics are a powerful tool to identify phenological
patterns and characterize the influence of the mixed layer.
• The relevance of metrics ΔINIT and ΔMAX was
emphasized.
• In the Mediterranean Sea, we proposed some hypotheses to
explain their behaviors. (Lavigne et al. JGR, in revision)
• The proposed phenological metrics could be particularly
adapted for profiling floats observations.
Thanks to
Fabrizio D’Ortenzio (my supervisor),
Loïc Houpert (CEFREM, Perpignan FRANCE) and
Rosario Lavezza (SZN, Napoli, Italy) for their help on CTD and
nitrate data.
Thank you to for your attention
Contact: [email protected]
Method Lavigne et al., 2012

similar documents