Approach: Assimilation Efficiencies

1
Global Phytoplankton Assimilation
Efficiencies from Carbon Based Productivity
Estimates
Robert T. OMalley, Michael J. Behrenfeld, Toby
K. Westberry Oregon State University, Department
of Botany and Plant Pathology, Corvallis, Oregon

• Approach Assimilation Efficiencies
• The Carbon based model calculates mixed layer NPP
  (mg m-3) as a function of carbon and
  phytoplankton growth rate
• NPP_mixed layer C u
• Rearranging the chlorophyll-based approach and
  implementing CbPM NPP data allows pb_opt to be
  calculated as
• pb_opt NPP_mixed layer / (chl day length)
• The resulting global patterns are shown below in
  Figure 2.
• These pb_opt values can then be compared to the
  two most common temperature-dependent functions
• The standard VGPM uses a polynomial function of
  sea surface temperature (SST) to model pb_opt
  based on a fit to empirical data. pb_opt is
  maximum for this model at 20 degrees C.
• Global patterns for the polynomial are shown in
  Figure 3.
• The Eppley uses an exponential function of SST
  to model pb_opt, normalized to go through 4.6 at
  20 degrees C.
• Global patterns for the Eppley curve are
  shown in Figure 4.

Introduction Net primary production (NPP) is a
key attribute of marine ecosystems. A wide range
of approaches have been developed to estimate
global distributions of NPP. One of the most
common approaches is to calculate NPP as a
function of satellite chlorophyll concentration
(chl), an hourly carbon assimilation efficiency
(pb_opt), and a volume function NPP Chl
pb_opt day length volume function A new
alternative approach is the Carbon-based
Production Model (CbPM) (Behrenfeld et al. 2005
GBC). The CbPM has recently been refined to
resolve vertical characteristics of water column
photosynthesis (see poster by Toby Westberry
Carbon-based primary production phytoplankton
physiology from ocean color data). NASAs
MODIS program is supporting the construction of a
productivity website for supplying
satellite-based NPP estimates, model code, and
supporting data http//www.science.oregonstate.
edu/ocean.productivity This single source of
internally consistent NPP data sets makes
comparisons between model predictions
straightforward. A primary source
of discrepancy between chlorophyll-based NPP
models is the description of pb_opt. Temperature
is the most common predictor of pb_opt in these
models, although such relationships are
correlative rather than causative. The new CbPM
model provides a means for calculating pb_opt
from satellite ocean color data. Here we use
data and code available through the productivity
web site to investigate relationships between
satellite based pb_opt values and predictions
based on earlier temperature-dependent models.

• Results and Observations
• CbPM-based scatter diagrams of pb_opt vs
  temperature show both remarkable variability and
  organized structure, as demonstrated by boreal
  winter and boreal summer data for 2003 (Figure
  5).
• Asymmetry in the relationship between pb_opt
  and temperature is striking between the northern
  and southern hemispheres for a given season
  (compare Figure 5A with 5B, and Figure 5C with
  5D).
• The boreal winter pb_opt-vs-temperature
  distribution in the northern hemisphere is
  consistent with the boreal summer distribution in
  the south (and vice versa) (compare Figure 5A
  with 5D, and Figure 5B with 5C). The annual
  month-to-month distributions progress smoothly
  between these extremes.
• By overlaying data from the Pacific basin in
  blue, and the Atlantic basin in brown, we see
  variations in structure between these basins,
  while spatially coherent structure can be seen
  within each basin itself. This structure can be
  seen in the parabolic shapes and arcuate
  lineations within the scatter diagrams.
• For comparison, the two temperature-dependent
  pb_opt functions (see Approach) are also overlain
  as red and blue curves in Figure 5.
• Interesting Features There is always a peak to
  the data, similar to the maximum in the VGPM
  curve, although it is at 25 degrees C, instead
  of 20 degrees C. At higher temperatures it never
  continues upward exponentially like the Eppley
  curve.

N
S
winter
winter
5A
5B
Points drawn in order of all data (black)
Pacific data (cyan) Atlantic data (brown). The
VGPM pb_opt curve is overlaid in red the Eppley
pb_opt curve is drawn in blue.
N
S
summer
summer
5C
5D
Figure 5

• Notes
• The carbon model that Toby Westberry is working
  on is a refinement of the initial model developed
  by Mike Behrenfeld. The new model allows for
  better resolution of NPP below the mixed layer.
  Westberry presented initial results at the Ocean
  Sciences meeting in 2006.
• The VGPM pb_opt function is a 7th order
  polynomial function of SST, in the form of
  pb_opt sum(i0,7) ai (sst/10)i
• The Eppley pb_opt function is scaled to fit 4.6
  at 20 degrees C after Antoine and Morel (1996)
  but left in the form Eppley (1972) originally
  gave his maximum growth function pb_opt 1.54
  10((0.0275sst) 0.07)

Figure 2
Figure 3
Figure 4
Figure 1
Contact Information Robert OMalley
omalleyr_at_science.oregonstate.edu Mike Behrenfeld
mjb_at_science.oregonstate.edu Toby Westberry
westbert_at_science.oregonstate.edu For NPP data,
code, model comparisons, field data, etc, please
see http//www.science.oregonstate.edu/ocean.prod
uctivity
PB_opt calculated using an exponential function
of sst. Max pb_opt values found at maximum sst
locations, causing the pattern of equatorial
highs seen here. Small global variations are
seen between winter (January) and summer (July).
Pb_opt calculated using a polynomial function of
sst AVHRR bsst files for January and July,
2003, were used for input. . Max pb_opt values
are at 20 degrees C, causing the mid-latitude
highs seen here. Small global variations are
seen between winter (January) and summer (July).
Pb_opt calculated using the spectral Carbon model
of Westberry. Max pb_opt observed in January off
the coast of Central America in July it is
seen to the west of South America. A high degree
of variation is seen from January to July, along
with variation between the northern and southern
hemispheres on those months. See the scatter
diagrams above.