Carbon-based primary production

1
Carbon-based primary production phytoplankton
physiology from ocean color data Toby Westberry1,
Mike Behrenfeld1, Emmanuel Boss2, Dave Siegel3,
Allen Milligan1 1Department of Botany Plant
Pathology, Oregon State University, 2School of
Marine Sciences, University of Maine, 3Institute
for Computational Earth System Science,
University of California Santa Barbara
1. Introduction
2. Model
Historically, net primary production (NPP) has
been modeled as a function of chlorophyll
concentration, allowing for a natural application
to ocean color satellite data. However, cellular
chlorophyll content is highly variable and is
affected by photoacclimation and nutrient stress
which act to confound global NPP model results.
An approach to alleviate these limitations was
recently introduced providing satellite NPP
estimates based on conversion of backscattering
to phytoplankton carbon and mixed layer
phytoplankton growth rates (m, day-1) derived
from chlorophyllcarbon ratios (Carbon Based
Productivity Model (CBPM), see Behrenfeld et al.,
2005).  Here, the CBPM is extended to provide
full vertical profiles of m and NPP given
knowledge of the mixed layer depth and the
nitracline depth, which allow appropriate
parameterization of photoacclimation and nutrient
stress through the water column. This depth
resolved approach accurately reconstructs the
underwater light field, and produces biological
profile data (C, Chl, NPP) which are broadly
consistent with laboratory- and field-based
measurements. Direct validation using regional in
situ datasets of phytoplankton chlorophyllcarbon,
cellular growth rates, and measured NPP rates
support the findings presented here.
Conceptual model (ex., strongly stratified case)
Some details
- within ML, phytoplankton are acclimated to
median PAR in ML. Below this, they are
acclimated to ambient PAR - Kd(490) is expanded
to give Kd(l) using Austin Petzold (1986) -
Kd(l,z) below ML is estimated from Chl
(Morel, 1988) difference between this
method Austin Petzold (1986) at surface -
each property (Kd,PAR,Chl,C,m,NPP) is dependent
on values above it (iterative!)
z0
Uniform
zMLD
Propagation (and relaxation) of surface nutrient
stress
Photoacclimated ChlC
Nutrient-limited /or light-limited
photoacclimation
y0
zzNO3
Light-limited photoacclimation
- Concept of a non-zero minimum ChlC (y0)
when m0 is important! - can be seen in in situ
data (e.g., Laws Bannister
(1980)) - required to make model formulations
self-consistent
Light limitation
Nutrient stress
z8
Relative PAR Relative NO3
2 main ingredients for modeling NPP 1) biomass
2) physiology
NPP biomass x physiologic rate
Can have any combination of above processes
depending on PAR(z) and depths of mixed layer and
nitracline
General
Most often the case

• Key Points
• Invert ocean color data to estimate Chl
  bbp(443)
• (Garver Siegel, 1997 Maritorena
  et al., 2001)
• Relate bbp(443) to carbon biomass (mg C m-2)
• Use satellite ChlC to infer phytoplankton
  physiology in mixed layer
• (photoacclimation nutrient stress)
• Iteratively propagate properties below mixed
  layer as ?(PAR,zMLD, zNO3)
• keeping track of photoacclimation to variable
  nutrient and light stress

Data Sources implementation
NPP Chl x Pbopt x
(e.g., Antoine Morel, 1996 Behrenfeld
Falkowski, 1997 others)
Intercept represents stable component of bbp due
to background particles (i.e., bacteria,
colloids, detritus, etc.) and scalar was chosen
such that phytoplankton C 30 of
scattering-based POC estimates
Chl-based
bbp (m-1)
INPUTS
Problem Chl does not adequately characterize
biomass for modeling NPP due to
photoacclimation and nutrient-
stress related changes in cellular chlorophyll
- SeaWiFS nLw(l)
Chl, bbp(443) PAR
Kd(490)
Kd(l) - FNMOC MLD - WOA 2001 NO3
ZNO3
OUTPUTS
Maritorena et al. (2001)
- Chl(z), C(z), ChlC(z) - m(z) - NPP(z) -
Kd(l,z), PAR(z), Zeu?
Austin Petzold (1986)
Chl (mg m-3)
Solution estimate biomass independently of
physiology
NPP C x m
Can estimate these from ocean color measurements
C-based
?NO3gt0.5mM
Chl C
m (divisions d-1)
1 x1 monthly mean climatologies (1999-2004)
Scattering (cp or bbp)
Ratio of Chl to scattering (ChlC)
Carbon-Based Production Model (CBPM)
Ig (Ein m-2 h-1)
3. Data Results
Depth-integrated NPP temporal patterns
Phytoplankton growth rates, mz0
Summer (Jun-Aug)
Example, single station (pixel
Ex 2. Tropical Atlantic
Ex. 1. North Atlantic
Western N. Atl
Eastern Pacific (20N, -110E, Aug)
All data
Oligotrophic gyres (L0)
CBPM

• In N. Atlantic, both the
• onset and peak of the
• spring bloom occur 1-2
• months later than Chl-
• based model indicates
• Seasonality is amplified at
• higher latitudes dampened
• in tropics (difference btwn
• summer winter)

VGPM

• Subsurface profile features
• variable Kd(l,z)
• subsurface Chl a maximum
• realistic 1 light levels
• variable C biomass
• realistic NPP profiles

• m represents a net community
• growth rate and reflects growth,
• respiration, and other losses
• (grazing, mixing, etc.) of a mixed
• population
• Surface phytoplankton growth
• rates are broadly consistent with
• expected values
• Median values of m are 0.4
• Open ocean growth rates range
• from 0.1-1.0 divisions d-1

Depth (m)
Eastern N. Atl
Winter (Dec-Feb)
m (d-1)
m (d-1)
occur.
CBPM
VGPM
Summary/Conclusions
Mean NPP profiles

• We have developed a spectral- depth-resolved
  NPP model based on independent C Chl estimates
  from ocean color data
• Ability to distinguish DChl due to
  photoacclimation v. growth
• Estimates of the phytoplankton growth rate, m,
  are also derived as part of this approach and
  provide valuable insight into phytoplankton
  physiology
• Spatial temporal patterns in NPP are markedly
  different than if using a Chl-based model (e.g.,
  VGPM of Behrenfeld Falkowski, 1997)
• Ongoing validation with various diagnostics
  (PAR, Chl, m, NPP) suggest the model is
  performing well

m (d-1)
Validation example NPP at Hawaii Ocean
Time-series
Depth-integrated NPP spatial patterns
Summer (Jun-Aug)
Summer (Jun-Aug)

• Large spatial differences between C-based
• and Chl-based NPP estimates (e.g., N.
• Atlantic, Eq. upwelling) lead to redistribution
  
• of NPP in time and space

If classify pixels according to average seasonal
Chl variance (see above image), such that
L0 L4 Oligotrophic High
Lat. we can look at mean m NPP
profiles within those regions
Depth (m)
?NPP VGPM This model
(Pg C yr-1)
Annual 45 49
Gyres 5 (11) 10 (20)
High latitudes 19 (42) 13 (27)
Subtropics 18 (40) 23 (47)
Southern Ocean (qlt-50S) 2 (4) 3 (6)
Depth (m)
Winter (Dec-Feb)
Winter (Dec-Feb)
NPP (mg C m-3 d-1)
Annual NPP (and of total) in each region
Chl-based NPP (mg C m-2 d-1)
C-based NPP (mg C m-2 d-1)