ArticlePDF Available

Quantifying spatiotemporal variability in zooplankton dynamics in the Gulf of Mexico with a physical–biogeochemical model


Abstract and Figures

Zooplankton play an important role in global biogeochemistry, and their secondary production supports valuable fisheries of the world's oceans. Currently, zooplankton standing stocks cannot be estimated using remote sensing techniques. Hence, coupled physical–biogeochemical models (PBMs) provide an important tool for studying zooplankton on regional and global scales. However, evaluating the accuracy of zooplankton biomass estimates from PBMs has been a major challenge due to sparse observations. In this study, we configure a PBM for the Gulf of Mexico (GoM) from 1993 to 2012 and validate the model against an extensive combination of biomass and rate measurements. Spatial variability in a multidecadal database of mesozooplankton biomass for the northern GoM is well resolved by the model with a statistically significant (p
Content may be subject to copyright.
Biogeosciences, 17, 3385–3407, 2020
© Author(s) 2020. This work is distributed under
the Creative Commons Attribution 4.0 License.
Quantifying spatiotemporal variability in zooplankton dynamics in
the Gulf of Mexico with a physical–biogeochemical model
Taylor A. Shropshire1,2, Steven L. Morey3, Eric P. Chassignet1,2, Alexandra Bozec1,2, Victoria J. Coles4,
Michael R. Landry5, Rasmus Swalethorp5, Glenn Zapfe6, and Michael R. Stukel1,2
1Department of Earth Ocean and Atmospheric Sciences, Florida State University, Tallahassee, FL 32303, USA
2Center for Ocean-Atmospheric Prediction Studies, Florida State University, Tallahassee, FL, USA
3School of the Environment, Florida A&M University, Tallahassee, FL, USA
4University of Maryland Center for Environmental Science, PO Box 775, Cambridge, MD 21613, USA
5Integrative Oceanography Division, Scripps Institution of Oceanography, 8622 Kennel Way, La Jolla, CA 92037, USA
6University of Southern Mississippi, Division of Coastal Sciences, Hattiesburg, MS 39406, USA
Correspondence: Taylor A. Shropshire (
Received: 25 November 2019 – Discussion started: 9 December 2019
Revised: 27 April 2020 – Accepted: 11 May 2020 – Published: 5 July 2020
Abstract. Zooplankton play an important role in global bio-
geochemistry, and their secondary production supports valu-
able fisheries of the world’s oceans. Currently, zooplankton
standing stocks cannot be estimated using remote sensing
techniques. Hence, coupled physical–biogeochemical mod-
els (PBMs) provide an important tool for studying zooplank-
ton on regional and global scales. However, evaluating the
accuracy of zooplankton biomass estimates from PBMs has
been a major challenge due to sparse observations. In this
study, we configure a PBM for the Gulf of Mexico (GoM)
from 1993 to 2012 and validate the model against an exten-
sive combination of biomass and rate measurements. Spa-
tial variability in a multidecadal database of mesozooplank-
ton biomass for the northern GoM is well resolved by the
model with a statistically significant (p< 0.01) correlation
of 0.90. Mesozooplankton secondary production for the re-
gion averaged 66 ±8×109kg C yr1, equivalent to 10 %
of net primary production (NPP), and ranged from 51 to
82 ×109kg C yr1, with higher secondary production in-
side cyclonic eddies and substantially reduced secondary
production in anticyclonic eddies. Model results from the
shelf regions suggest that herbivory is the dominant feeding
mode for small mesozooplankton (< 1 mm), whereas larger
mesozooplankton are primarily carnivorous. In open-ocean
oligotrophic waters, however, both mesozooplankton groups
show proportionally greater reliance on heterotrophic protists
as a food source. This highlights an important role of mi-
crobial and protistan food webs in sustaining mesozooplank-
ton biomass in the GoM, which serves as the primary food
source for early life stages of many commercially important
fish species, including tuna.
1 Introduction
Within marine pelagic ecosystems, zooplankton function as
an important energy pathway between the base of the food
chain and higher trophic levels such as fish, birds and mam-
mals (Landry et al., 2019; Mitra et al., 2014). Zooplank-
ton also have a well-documented impact on chemical cy-
cling in the ocean (Buitenhuis et al., 2006; Steinberg and
Landry, 2017; Turner, 2015). The ecological roles of zoo-
plankton, however, are varied and taxon dependent. Glob-
ally, protistan grazing is the largest source of phytoplankton
mortality, accounting for 67% of daily phytoplankton growth
(Landry and Calbet, 2004). Protistan zooplankton function
primarily within the microbial loop, leading to efficient nu-
trient regeneration in the surface ocean (Sherr and Sherr,
2002; Strom et al., 1997). By contrast, mesozooplankton con-
tribute significantly less to phytoplankton grazing pressure,
consuming an estimated 12 % of primary production glob-
ally (Calbet, 2001), but strongly impact the biological car-
bon pump. In addition to grazing pressure on phytoplankton,
mesozooplankton affect the biological carbon pump through
Published by Copernicus Publications on behalf of the European Geosciences Union.
3386 T. A. Shropshire et al.: Quantifying spatiotemporal variability in zooplankton dynamics
top-down pressure on protistan grazers, production of sink-
ing fecal pellets, consumption of sinking particles, and active
carbon transport during diel vertical migration (Steinberg and
Landry, 2017; Turner, 2015). Herbivorous mesozooplankton
are particularly important to study as they are often asso-
ciated with shorter food chains that enable efficient energy
transfer from primary producers to higher trophic levels of
immediate societal interest such as economically valuable
fish species and/or their planktonic larvae.
Zooplankton populations have been identified as being
vulnerable to impacts of a warming ocean (Caron and
Hutchins, 2013; Pörtner and Farrell, 2008; Straile, 1997),
through direct temperature effects on metabolic rates (Ikeda
et al., 2001; Kjellerup et al., 2012) and thermal stratification-
driven alterations in food web structure and productivity
(Landry et al., 2019; Richardson, 2008). Studies aimed
at monitoring and predicting zooplankton populations are
therefore critical for understanding the first-order effects of a
warming ocean on marine ecosystems given the importance
of secondary production and the impact zooplankton have on
biogeochemical cycling. Despite their importance, zooplank-
ton have been historically sampled with limited temporal
and spatial resolution. Unlike ocean hydrodynamics and phy-
toplankton variability, zooplankton abundance cannot cur-
rently be estimated remotely from space. Thus, numerical
models provide a useful tool for synoptic assessments of zoo-
plankton stocks on basin and global scales (Buitenhuis et al.,
2006; Sailley et al., 2013; Werner et al., 2007). Nonethe-
less, evaluating the accuracy of zooplankton abundance es-
timates in numerical experiments, such as three-dimensional
physical–biogeochemical ocean models (PBMs), is a major
challenge due to the sparse ship-based observations in most
regions (Everett et al., 2017). Consequently, PBMs are typ-
ically predominately validated against surface chlorophyll
(Chl) from remote sensing products (Doney et al., 2009;
Gregg et al., 2003; Xue et al., 2013).
In most marine environments, phytoplankton net growth
rates and biomass are determined primarily by the imbal-
ance between phytoplankton growth and zooplankton graz-
ing (Landry et al., 2009). PBMs can accurately predict phy-
toplankton standing stock (i.e., compare well with satellite
Chl observations) despite being driven by the wrong under-
lying dynamics, leading to major errors in model estimates of
secondary production and nutrient cycling (Anderson, 2005;
Franks, 2009). For instance, parameter tuning using only sur-
face Chl as a validation metric can allow broad patterns in
phytoplankton biomass to be reproduced even with gross
over- or underestimation of phytoplankton turnover times.
Similarly, even a model that is validated against satellite Chl
and net primary production might completely misrepresent
the proportion of phytoplankton mortality mediated by zoo-
plankton groups, leading to inaccurate estimates of important
ecological metrics like secondary production and carbon ex-
port. Hence, validating PBMs against zooplankton dynam-
ics is key to increasing confidence in model solutions. The
importance of validation is further evident when consider-
ing zooplankton impacts on the behaviors of biogeochemi-
cal models (Everett et al., 2017). Differences in simulated
zooplankton communities expressed through the number of
functional types, various mathematical grazing functions, or
the arrangement of transfer linkages have been shown to have
substantial impacts on the dynamics of simple and complex
biogeochemical models (Gentleman et al., 2003b; Gentle-
man and Neuheimer, 2008; Mitra et al., 2014; Murray and
Parslow, 1999; Sailley et al., 2013).
The Gulf of Mexico (GoM) is a particularly suitable
region for examining zooplankton dynamics with PBMs.
In the northern and central Gulf of Mexico, zooplankton
abundances have been extensively measured for over three
decades (1982–present) by the Southeast Area Monitoring
and Assessment Program (SEAMAP). Within the SEAMAP
dataset, zooplankton biomass exhibits strong spatiotempo-
ral variability, reflecting complex physical circulation in the
GoM. Circulation off the shelf is characterized by sub-
stantial upper-layer mesoscale activity driven primarily by
the energetic Loop Current (Forristall et al., 1992; Maul
and Vukovich, 1993; Oey et al., 2005). In contrast, coastal
and shelf circulation patterns are predominantly wind-driven
(Morey et al., 2003a, 2013). Freshwater discharged by the
Mississippi River and other smaller rivers is frequently en-
trained offshore by shelf break interaction with mesoscale
features (e.g., anticyclonic Loop Current eddies), leading to
strong horizontal and vertical gradients in physical and bio-
geochemical quantities (Morey et al., 2003b). Overlap of
these gradients with the SEAMAP study region results in
zooplankton collections across biogeochemically heteroge-
neous environments, providing a powerful model constraint.
For instance, Chl ranges across 3 orders of magnitude (
0.01–10 mg Chl m3) from oligotrophic to eutrophic waters.
Over the past decade several PBM studies have been con-
ducted in the GoM, all primarily examining nutrient and phy-
toplankton dynamics. Early work by Fennel et al. (2011) ex-
amined phytoplankton dynamics on the Louisiana and Texas
continental shelf, concluding that loss terms (e.g., grazing)
rather than growth rates dictated accumulation rates of phy-
toplankton biomass. With the same biogeochemical model,
Xue et al. (2013) conducted the first gulf-wide PBM study to
investigate broad seasonal biogeochemical variability and to
constrain a shelf nitrogen budget. More recently, Gomez et
al. (2018) implemented a biogeochemical model with multi-
ple phytoplankton and zooplankton functional types to gain a
more detailed understanding of nutrient limitation and phy-
toplankton dynamics in the GoM. To examine phytoplank-
ton seasonality and biogeography in the oligotrophic Gulf of
Mexico, Damien et al. (2018) validated a PBM based on a
unique subsurface autonomous glider dataset. Together, these
studies have demonstrated the utility of PBMs for investigat-
ing GoM lower trophic levels and have also highlighted the
key ecosystem roles of zooplankton. Specifically, both Fen-
nel et al. (2011) and Gomez et al. (2018) identified the impor-
Biogeosciences, 17, 3385–3407, 2020
T. A. Shropshire et al.: Quantifying spatiotemporal variability in zooplankton dynamics 3387
tance of zooplankton in modulating the simulated seasonal
patterns of phytoplankton biomass, emphasizing the impor-
tance of top-down control on the shelf. Although simulated
zooplankton community results were not presented, Damien
et al. (2018) noted that biotic processes such as grazing pres-
sure are “essential to fully understanding the functioning of
the GoM ecosystem”. However, in all of these studies, zoo-
plankton validation was largely absent.
In this study, we configured a PBM for the GoM to esti-
mate zooplankton abundance and analyze zooplankton com-
munity dynamics. The PBM is forced by three-dimensional
hydrodynamic fields from a data assimilative HYbrid Coor-
dinate Ocean Model (HYCOM) hindcast of the GoM (http:
//, last access: 29 June 2020) and is based on
the biogeochemical model NEMURO (North Pacific Ecosys-
tem Model for Understanding Regional Oceanography; Kishi
et al., 2007), which is substantially modified here for appli-
cation to the GoM. The model is integrated over 20 years
(1993–2012) and validated against an extensive combination
of remote and in situ measurements including total and size-
fractioned mesozooplankton biomass and grazing rates, mi-
crozooplankton grazing rates, phytoplankton growth rates,
and net primary production as well as validation of surface
chlorophyll and vertical profiles of chlorophyll and nitrate.
Our goals were (1) to develop and validate a PBM to es-
timate mesozooplankton abundance, (2) to characterize the
spatiotemporal variability in mesozooplankton dietary com-
position, and (3) to quantify regional mesozooplankton sec-
ondary production. We focus primarily on the oligotrophic
open-ocean GoM, where prey (i.e., zooplankton) availability
may be limiting for fish, their larvae and other higher trophic
2 Methods and data
2.1 Biogeochemical model configuration
2.1.1 NEMURO model description
The biogeochemical model for this study is based on NE-
MURO (Kishi et al., 2007) but has been modified and
parameterized to more accurately reflect the ecology of
the GoM (herein called NEMURO-GoM). NEMURO is a
concentration-based, lower-trophic-level ecosystem model
originally developed and parameterized for the North Pacific.
Like most marine functional-group biogeochemical models,
it is structured around simplified representations of the lower
food web originating from earlier nutrient–phytoplankton–
zooplankton models (Fasham et al., 1990; Franks, 2002; Ri-
ley, 1946). Complexity is added through additional state vari-
ables and transfer functions with the specific goal of resolv-
ing dynamics within the nutrient, phytoplankton and zoo-
plankton pools. In total, NEMURO has 11 state variables: six
nonliving state variables – nitrate (NO3), ammonium (NH4),
dissolved organic nitrogen (DON), particulate organic ni-
trogen (PON), silicic acid (Si(OH)4) and particulate silica
(opal); two phytoplankton state variables – small (SP) and
large phytoplankton (LP); and three zooplankton state vari-
ables – small (SZ), large (LZ) and predatory zooplankton
Each biological state variable in NEMURO is an ag-
gregated representation of taxonomically diverse plankton
groups that function similarly in the ecosystem. The phyto-
plankton community is modeled as two functional types of
obligate autotrophs: small phytoplankton (SP, predominantly
cyanobacteria and picoeukaryotes in the GoM) and large
phytoplankton (LP, diatoms). Small zooplankton (SZ) rep-
resent heterotrophic protists, and metazoan zooplankton are
divided into suspension-feeding mesozooplankton (LZ) and
predatory zooplankton (PZ). Here, we assume that LZ and
PZ are nonmigratory. Heterotrophic bacteria are implicitly
represented by temperature-dependent decomposition rates,
which represent nitrification and remineralization processes.
NEMURO uses nitrogen as a model currency since it is the
major limiting macronutrient in much of the ocean. Silica is
also included as a potentially colimiting nutrient for diatoms
(i.e., LP).
By default, sinking in NEMURO is restricted to PON
and opal, and benthic processes are not included. Here, be-
cause of the large shelf area in the GoM, we implemented
a simple diagenesis of PON/opal to NO3/SiO4and removal
of PON/opal through sedimentation, where 1 % of the flux
sinking out of bottom cell was removed and 10 % converted
back into NO3/SiO4. However, we found that this had no
significant impact on the simulated surface Chl or meso-
zooplankton biomass on the shelf. The inclusion of a more
complex sediment diagenesis model (including denitrifica-
tion) would have added further realism (Fennel et al., 2011).
However, our main focus was to evaluate zooplankton dy-
namics in the oligotrophic region where higher trophic levels
that depend on mesozooplankton secondary production may
experience food limitation and where benthic processes are
NEMURO was chosen for the present study because it dis-
tinguishes SZ, LZ and PZ, permitting a detailed analysis of
dynamics for multiple functional types in the GoM zooplank-
ton community. During initial GoM simulations, default NE-
MURO parameterizations, configured for the North Pacific
(Kishi et al., 2007), substantially overestimated surface ni-
trate, surface Chl and mesozooplankton biomass relative to
observations. We attribute these differences to (1) substan-
tially higher temperatures in the GoM compared with the
North Pacific, which significantly increase decomposition
and growth rates in the model, resulting in higher nutrient
recycling and elevated near-surface stocks of phytoplankton
and zooplankton; and (2) distinct differences in taxonomic
composition of phytoplankton and zooplankton communities
in the GoM and North Pacific, with significant differences in
key parameter values associated with growth and grazing. Biogeosciences, 17, 3385–3407, 2020
3388 T. A. Shropshire et al.: Quantifying spatiotemporal variability in zooplankton dynamics
For more details on the specific processes represented and
the interactions between state variables in NEMURO, we di-
rect readers to Kishi et al. (2007). All model equations are
provided in the Supplement to this paper. Biogeochemical
model forcing, initial and open boundary conditions are also
outlined in Supplement Sect. S1. Briefly, daily-average short-
wave radiation fields obtained from Climate Forecast Sys-
tem Reanalysis (CFSR) were used to force light limitation
of phytoplankton. Once a final parameter set was determined
(see Sect. 2.1.3), initial and open boundary conditions for all
state variables were prescribed from a spun up idealized one-
dimensional version of NEMURO-GoM. After initializing,
the three-dimensional model was spun up over four years be-
fore conducting the full 20-year experiment. River nutrient
input from the Mississippi River was prescribed using ni-
trate samples collected by United States Geological Survey
(USGS) and due to a lack of observations for other rivers was
prescribed for all 37 rivers represented in the model.
2.1.2 Modifications to default NEMURO model
To improve realism for application to the GoM, five struc-
tural changes were made to the original NEMURO model.
First, we removed the SP to LZ grazing pathway. The original
SP state variable for the North Pacific represents nanophy-
toplankton (e.g., coccolithophores), which can be important
prey of copepods and other mesozooplankton. In the GoM,
however, cyanobacteria and picoeukaryotes (too small for di-
rect feeding by most mesozooplankton) comprise much of
the phytoplankton biomass and hence are represented as SP
in our model. In addition to adding ecological realism, this
change in direct trophic connection between SP and LZ al-
lowed the model to produce a more realistic LP-dominated
phytoplankton community on the shelf (see Discussion).
Next, quadratic mortality was replaced with linear mor-
tality for all biological state variables with the exception
of predatory zooplankton (PZ). In biogeochemical mod-
els, quadratic mortality is often used for numerical stabil-
ity and/or to represent implicit loss terms to an unmodeled
parasite or predator that covaries in abundance with its prey
(e.g., viral lysis of phytoplankton or predation by unmodeled
higher predators) (Anderson et al., 2015). However, grazing
mortality is explicitly modeled in NEMURO, and viral mor-
tality is generally not a substantial loss term for bulk phy-
toplankton (Staniewski and Short, 2018). Quadratic mortal-
ity was retained for PZ to account for predation pressure of
unmodeled predators (e.g., planktivorous fish). During the
model tuning process, we found that removal of quadratic
mortality from the four other plankton functional groups was
an important parameterization change that allowed the model
to simulate more realistic mesozooplankton biomass in the
oligotrophic GoM (see Discussion).
The default ammonium inhibition term and light limitation
functional form in NEMURO were replaced in NEMURO-
GoM with more widely adopted parameterizations. The ex-
ponential ammonium inhibition term in the nitrate limi-
tation function was replaced with the term described by
Parker (1993), as has been done in previous PBM studies
(Fennel et al., 2006) due to the nonmonotonic behavior of
the default NEMURO ammonium inhibition term. At high
NO3concentrations, the default term is known to generate
unrealistic phytoplankton nutrient uptake patterns in which
total nutrient uptake (i.e., uptake of NO3plus uptake of NH4)
can actually decrease despite increases in NH4(and constant
Light limitation in NEMURO is based on an optimal light
parameterization that implicitly includes photoinhibition.
This formulation was replaced with the Platt et al. (1980)
functional form that allows one to explicitly control the
amount of photoinhibition, which can be important in the
GoM where surface irradiances are high. Additionally, the
Platt functional form is commonly used, and thus parame-
ter values are easier to find for comparison (e.g., initial slope
of the PI curve, α). This formulation is also implemented in
newer versions of NEMURO, such as the code used in the
Regional Ocean Modeling System (ROMS) NEMURO bio-
geochemical package.
Finally, to account for photoacclimation and more accu-
rately simulate deep chlorophyll maximum (DCM) dynam-
ics, we replaced the constant C :Chl parameter with a vari-
able C :Chl model where ratios for SP and LP were allowed
to vary based on the formulation described by Li et al. (2010),
which considers both light and nutrient limitation (see Sup-
plement). The Li et al. (2010) equations build on a previ-
ously constructed dynamic regulatory model of phytoplank-
ton physiology which describes C :Chl variability under bal-
anced growth and nutrient-saturated conditions at constant
temperature (see Geider et al., 1998). Herein, “default” NE-
MURO includes the modified ammonium inhibition, light
formulation and variable C :Chl model.
2.1.3 NEMURO-GoM model tuning procedure
In total, NEMURO includes 71 parameters, 23 of which
were modified in the present study. For initial model tun-
ing, we used an idealized one-dimensional model designed
to mimic the oligotrophic GoM. To guide our tuning pro-
cedure, we relied on a semiquantitative approach where the
one-dimensional model solution was evaluated based on five
ecosystem benchmarks. Target values for benchmarks and
other ecosystem attributes were determined from observa-
tions or a theoretical basis. Ecosystem benchmarks included
surface Chl, mesozooplankton biomass, DCM depth, DCM
magnitude, and SP :LP ratio. Surface Chl and mesozoo-
plankton biomass were chosen as benchmarks to evaluate the
realism of plankton biomass in the model. The DCM depth
and magnitude were chosen to evaluate the vertical structure
of the simulated ecosystem, and the SP :LP ratio was used
to gauge the realism of the plankton community composition
(i.e., high SP :LP is expected in the oligotrophic GoM). The
Biogeosciences, 17, 3385–3407, 2020
T. A. Shropshire et al.: Quantifying spatiotemporal variability in zooplankton dynamics 3389
model was also tuned by considering the relative magnitudes
of loss terms for phytoplankton (grazing, mortality, respira-
tion and excretion), total protistan zooplankton grazing rel-
ative to mesozooplankton grazing, and surface and deep ni-
trate concentrations. We outline each parameter change, jus-
tification and the resulting impact on the ecosystem bench-
marks simulated by the idealized one-dimensional model in
Sect. S3. Where possible, we modified parameters in groups
so that relative changes were consistent throughout the model
(e.g., doubling all zooplankton mortality terms). After tun-
ing in the one-dimensional model, parameter sets were im-
plemented into the full three-dimensional model where ad-
ditional tuning was performed. Once a final parameter set
was determined we conducted a parameter sensitivity analy-
sis over 18 individual experiments to identify impacts of pa-
rameter changes from default NEMURO values (Sect. S5).
2.2 Physical model configuration
2.2.1 Description of the offline numerical environment
To run large numbers of three-dimensional simulations ef-
ficiently for basin-scale tuning, NEMURO-GoM was run
offline using the MITgcm offline tracer advection pack-
age. MITgcm was selected as it contains convenient pack-
ages for running offline simulations (McKinley et al., 2004).
That is, the dynamical equations of motion are not com-
puted during the NEMURO-GoM integration, but rather
the physical prognostic variables (i.e., temperature, salinity
and three-dimensional velocity fields) are prescribed from
daily-averaged flow fields saved from a previous hydrody-
namic model integration. This allows the recycled use of
flow fields, leaving only the tracer equations to be com-
puted. In the offline MITgcm package, the prognostic vari-
ables provide input to an advection scheme and mixing rou-
tine that conservatively handles offline advection and diffu-
sion of the biogeochemical tracer fields. MITgcm has many
options for linear and nonlinear advection schemes. Here
we use a third-order direct space–time flux-limiting scheme.
Sub-grid-scale mixing of the biogeochemical fields is han-
dled offline through the nonlocal K-profile parameteriza-
tion (KPP) package based on mixing schemes developed by
Large et al. (1994). For more information about the MIT-
gcm packages, we direct readers to the MITgcm manual
(, last access: 29 June 2020).
There are two main advantages to running PBMs in an of-
fline environment: (1) the momentum equations are not inte-
grated during the model run; and (2) the physical time step is
no longer bound by the dynamical Courant–Friedrichs–Lewy
(CFL) numerical stability criterion, which together signifi-
cantly reduces the computational cost. Instead, the stability
of the tracer advection scheme and timescales needed to re-
solve biological/physical processes of interest set the limits
on the time steps and prescription frequencies of flow fields.
When the physical time step is shorter than the flow field
prescription frequency, a simple linear interpolation of the
flow fields is performed between time steps. Offline simu-
lations of tracer advection have been found to closely re-
semble online runs (that is, computed together with the inte-
gration of the hydrodynamic model’s prognostic equations)
when the three-dimensional flow fields are prescribed at a
frequency that is at or below the inertial period (T=2π/f ,
TGoM > 24 h) for a region (Hill et al., 2005).
In the present study, the offline time step (30 min) is an
order of magnitude greater than the hydrodynamic model’s
(HYCOM-GoM, described in Sect. 2.2.2) baroclinic time
step (120 s). For reference, HYCOM-GoM required 76 d
to run to completion on 64 parallel cores. These time re-
quirements would increase considerably with the 11 addi-
tional biogeochemical tracers used in NEMURO. In con-
trast, NEMURO-GoM ran significantly faster, taking a total
of 50 h on 80 parallel cores. Offline models offer a valu-
able tool for integrating PBMs particularly as spatial reso-
lution and complexity in these models continue to increase
(e.g., DARWIN, Follows et al., 2007; GENOME, Coles et
al., 2017). While computationally advantageous, however,
offline simulations have inherently greater input and out-
put (I/O) demands that can become bottlenecks in some ap-
plications. Issues with conservation can also arise as three-
dimensional advection schemes are only approximately pos-
itive definite.
2.2.2 Description of the offline dynamical fields
The NEMURO-GoM model is forced by daily-averaged
three-dimensional velocity, temperature and salinity fields
from a preexisting 20-year (1993–2012) HYCOM (Chas-
signet et al., 2003) regional GoM hindcast (H-GoM). H-GoM
is based on version 2.2.99B of the HYCOM code, origi-
nally provided by the Naval Oceanographic Office (NAVO-
CEANO) Major Shared Resource Center. H-GoM was run at
1/25(4 km) horizontal resolution with 36 vertical hybrid
coordinate layers and assimilated historic, in situ and satel-
lite observations. The domain encompasses the entire GoM
and extends south of the Mexico—Cuba Yucatán Channel
to 18N and as far east as 77W (Fig. 1). Further details
on H-GoM (experiment ID: GOMu0.04/expt_50.1), includ-
ing model forcing and the main model configuration file (i.e.,
blkdat.input_501), can be found at
(last access: 29 June 2020).
The H-GoM flow fields were mapped from the HYCOM
native hybrid vertical coordinate to zlevels used by the MIT-
gcm. NEMURO-GoM was configured for 29 vertical zlev-
els (10 m intervals from 0 to 150 m; 25 m intervals from 150
to 300 m; 50 m intervals from 300 to 500; and 1000, 2000
and 4000 m). Mapping was performed by computing to-
tal zonal and meridional transports across the lateral bound-
aries of each MITgcm grid cell (e.g., 0–10 m bin; which may
include multiple HYCOM layers) and then dividing by the
area of the respective cell face. This vertical mapping ap- Biogeosciences, 17, 3385–3407, 2020
3390 T. A. Shropshire et al.: Quantifying spatiotemporal variability in zooplankton dynamics
proach is consistent as both HYCOM and MITgcm use an
Arakawa C-grid orientation for model variables. The H-GoM
bathymetry was adjusted such that no partial cells existed in
the domain to avoid thin cells. The continuity equation was
subsequently used to calculate vertical velocities. The use of
transports in this approach ensures conservation and approx-
imately identical profiles of vertical velocity to those in H-
GoM fields. For mapping of temperature and salinity fields
(used in the KPP mixing routine and for scaling biological
temperature-dependent rates), a simple linear interpolation
was performed.
2.3 Model validation
2.3.1 Surface chlorophyll observations
A benchmark for surface Chl was determined using the
Sea-Viewing Wide Field-of-View Sensor (SeaWiFS) prod-
uct from the Ocean Biology Processing Group (OBPG) of
the National Aeronautics and Space Administration (NASA).
The product used here is the mapped, level-3, daily, 9 km res-
olution images from 4 September 1997 to 10 December 2010
processed according to the algorithm of Hu et al. (2012).
To compute model–data point-to-point comparisons, we take
the corresponding daily-averaged simulated surface Chl field
and interpolate to the SeaWiFS grid before applying the
daily cloud coverage mask corresponding to the matching
SeaWiFS image. In total 4291 daily images consisting of
22 244 513 nonzero cell values (herein referred to SeaWiFS
measurements) were used to validate NEMURO-GoM. Ap-
proximately 500–1200 daily model–data point-to-point com-
parisons were made for each SeaWiFS grid cell (Fig. 1).
2.3.2 Mesozooplankton biomass observations
To evaluate model mesozooplankton biomass estimates, we
used plankton tows collected during Southeast Area Moni-
toring and Assessment Program (SEAMAP) surveys in the
northern and central GoM. In total, 11 781 plankton tows
were collected from 1983 to 2017, with two main annual
surveys in the spring (offshore) and fall (shelf) (Fig. 1). On
average, SEAMAP collected approximately 300 samples per
year with a specific sampling array offshore and more general
sampling coverage on the shelf. In total, 6835 samples were
used for direct point-to-point model–data comparisons. Sam-
ples were collected using standard gear consisting of a 61 cm
diameter bongo frame fitted with two 333 µm mesh nets. The
nets were fished in a double-oblique tow pattern from the
surface down to 200 or 5 m off the bottom and back to the
surface. Simultaneous samples were also collected using a
202 µm mesh net during 82 tows. Of these samples, roughly
half were collected in the oligotrophic GoM. The average ra-
tio between biomass measured in the 333 and 202 µm bongo
tows (0.5093 ±0.12) was used to convert 333 µm samples so
that direct comparisons could be made with model mesozoo-
plankton (LZ+PZ) biomass fields. Zooplankton biomass was
originally quantified as displacement volumes (DVs). Car-
bon mass (CM) equivalents were subsequently calculated as
log10(CM) =(log10 (DV) +1.434)/0.820 (Wiebe, 1988; Mo-
riarty and O’Brien, 2013).
For comparison to in situ biomass measurements, size
ranges were assumed for zooplankton state variables (see
Sect. 2.3.4). In NEMURO-GoM, SZ represent heterotrophic
protists (e.g., ciliates) and early stages of mesozooplankton,
which are typically < 200 µm in the ocean. Hence, SZ are
assumed to range in size from 2 to 200 µm. LZ are con-
sidered to represent small mesozooplankton ranging in size
from 0.2 to 1.0 mm, while PZ represent large mesozooplank-
ton ranging in size from 1.0 to 5.0 mm. For comparison to
the SEAMAP climatology the model LZ and PZ fields were
first converted to units of carbon assuming Redfield C :N
ratio and summed over the entire simulation before being
depth averaged to the bottom or 200 m. For point-to-point
model–data comparisons, total mesozooplankton biomass
fields were interpolated to SEAMAP sample locations/times
and then depth averaged to the corresponding sample tow
2.3.3 Observed vertical profiles of chlorophyll and
Depth profiles of Chl were also collected during SEAMAP
surveys using a Sea-Bird WETStar fluorometer attached to
a CTD. Calibration of the fluorimeter was infrequent, and
thus profiles were used to determine the depth of the flu-
orescence maxima for comparisons to DCM depths in the
model. In total, 2435 profiles were collected from 2003 to
2012, with 1052 profiles overlying bottom depths > 1000 m.
Profiles were available for earlier SEAMAP surveys; how-
ever, no standard QA/QC protocol for fluorometer data was
in place prior to 2003.
To evaluate DCM magnitudes in the model, we used 145
fluorescence profiles collected during May 2017 and 2018
process study cruises (see Sect. 2.3.4). The fluorometer was
attached to a CTD and calibrated using 126 in situ Chl
samples. Chl concentrations were determined from filtered
samples collected at depths ranging from 5 to 115 m using
high-performance liquid chromatography (HPLC). Since the
cruise sampling does not overlap with our NEMURO-GoM
simulation period, model–data comparisons were made for
all 20 years of the model run using sample locations and time
of the year. This was also done with other field measurements
from the process cruises (see Sect. 2.3.4).
For model–data comparisons of nitrate, we utilized pro-
files from the World Ocean Database (WOD). In total, 96
profiles were available during our simulation period and lo-
cated in the oligotrophic GoM (> 1000 m isobath). Profiles
were collected during all months except March and Decem-
ber, with the majority of samples collected during May, July
and August (Fig. 1a).
Biogeosciences, 17, 3385–3407, 2020
T. A. Shropshire et al.: Quantifying spatiotemporal variability in zooplankton dynamics 3391
Figure 1. (a–e) Spatial and temporal coverage of observational datasets used for model validation. Total number of nonzero values available
in daily SeaWiFS images from September 1997 to 10 December 2010 along with nitrate profiles from the World Ocean Database (black
dots) and the location of in situ measurements collected during May 2017 (circles) and May 2018 (triangles) process study cruises (a). Total
annual sampling of the SEAMAP surveys from 1983 to 2017 with samples overlapping with the PBM simulation period denoted in red (b).
SEAMAP samples organized by month (d). Total sample density within each 0.5×0.5box (c) along with the number of years with at least
one sample (e). The 1000 m isobaths and coastline are denoted by black continuous lines.
2.3.4 Biomass and rate measurements from process
study cruises
Although in situ rate measurements are made much less fre-
quently than biological standing stock measurements, they
offer very powerful constraints for validating the internal
dynamics of a biogeochemical model (Franks, 2009). Con-
sequently, we made phytoplankton and zooplankton rate
measurements on two cruises in the open-ocean GoM in
May 2017 and 2018 and used these measurements to vali-
date the model (Fig. 1a). On the process study cruises, we
utilized a quasi-Lagrangian sampling scheme to investigate
plankton dynamics in the oligotrophic GoM. Two drifting ar-
rays (one sediment trap array and one in situ incubation ar-
ray) were deployed to serve as a moving frame of reference
during 4 d studies (cycles) characterizing the water parcel
(Landry et al., 2009; Stukel et al., 2015). During these cycles,
we measured daily profiles of Chl, photosynthetically active
radiation, phytoplankton growth rates and productivity, pro-
tistan grazing rates, and size-fractionated mesozooplankton
biomass and grazing rates.
Size-fractionated mesozooplankton biomass and grazing
rates were determined from daily day–night paired oblique
ring-net tows (1 m diameter, 202µm mesh). In total, 40
oblique bongo net tows (16 in 2017 and 24 in 2018) sampled
the oligotrophic GoM mesozooplankton community from
near surface to a depth ranging from 100 to 135 m. Upon
recovery, the sample was anesthetized using carbonated wa-
ter, split using a Folsom splitter, filtered through a series
of nested sieves (5, 2, 1, 0.5 and 0.2 mm), filtered onto
preweighed 200 µm Nitex filters, rinsed with isotonic am-
monium formate to remove sea salt and flash frozen in liq-
uid nitrogen. In the lab, defrosted samples were weighed for
total wet weight and subsampled in duplicate (wet weight
removed) for gut fluorescence analyses. The remaining wet
sample was dried and subsequently reweighed and com-
busted for CHN analyses to determine total dry weight and
C and N biomasses. Gut fluorescence subsamples were ho-
mogenized using a sonicating tip, extracted in acetone, and
measured for Chl and phaeopigments using the acidification
method. The phaeopigment concentrations in the zooplank-
ton guts were the basis for calculated grazing rates using gut
turnover times based on temperature relationships for mixed
zooplankton assemblages. For additional details, see Décima
et al. (2011, 2016). Biogeosciences, 17, 3385–3407, 2020
3392 T. A. Shropshire et al.: Quantifying spatiotemporal variability in zooplankton dynamics
Protistan grazing rates were measured using the two-point,
mini-dilution variant of the microzooplankton grazing dilu-
tion method (Landry et al., 1984, 2008; Landry and Has-
sett, 1982). Briefly, one 2.8L polycarbonate bottle was gen-
tly filled with whole seawater taken from six depths (from the
surface to the depth of the mixed layer). A second 2.8 L bot-
tle was then filled with 33 % whole seawater and 67% 0.2µm
filtered seawater. Both bottles were then placed in mesh bags
and incubated in situ at natural depths for 24 h. These ex-
periments were conducted on each day of the 4 d cycle.
After 24 h, the bottles were retrieved and filtered onto glass
fiber filters, and Chl concentrations were determined using
the acidification method (Strickland and Parsons, 1972). Net
growth rates (k=ln(Chlfinal/Chlinit)) in each bottle were de-
termined relative to initial Chl samples. Phytoplankton spe-
cific mortality rates resulting from the grazing pressure of
protists were calculated as m=(kdk0)/(1–0.33), where kd
is the growth rate in the dilute bottle and k0is the growth
rate in the control bottle. Phytoplankton specific growth rates
were calculated as µ=k0+m. For additional details, see
Landry et al. (2016) and Selph et al. (2016). Phytoplankton
net primary production was quantified at the same depths by
3uptake experiments. Triplicate 2.8L polycarbonate
bottles and a fourth “dark” bottle were spiked with H13CO
and incubated in situ for 24 h at the same sampling depths
as for the dilution experiments. Samples were then filtered
and the 13C:12 C ratios of particulate matter determined by
isotope ratio mass spectrometry.
3 Results
3.1 Surface chlorophyll model–data comparisons
Model surface Chl estimates were found to agree closely
with satellite observations reproducing patterns in both the
oligotrophic and shelf region (Fig. 2). Spatial covariance be-
tween SeaWiFS climatology and model surface Chl clima-
tology (calculated with daily cloud cover mask applied) is
statistically significant (p< 0.01) with a correlation (ρ) of
0.72. When model estimates are compared to all 22 244 513
SeaWiFS measurements at corresponding times and loca-
tions (i.e., daily grid cell pairs), we find a ρvalue of 0.50
(p< 0.01). To facilitate more detailed model–data compar-
isons, the GoM domain was divided into an oligotrophic re-
gion (> 1000 m bottom depth) and a shelf region (< 1000 m
bottom depth). In the oligotrophic region, the correlation be-
tween model–data daily grid cell pairs is significant but weak
(ρ=0.17, p< 0.01) as a result of relatively low large-scale
spatial variability and hence dominance at the mesoscale.
However, bias is quite low (0.014 mg Chl m3), equivalent
to 10 % of the observed mean. In the shelf region, the cor-
relation is higher (ρ=0.47, p< 0.01) yet the bias is greater
(+0.90 mg Chl m3), equivalent to 92% of the mean. Previ-
ous GoM studies have determined ρvalues for monthly aver-
ages, which we calculate here for comparison. Based on 30 d
averages, the ρvalues are 0.70 (p<0.01) for the oligotrophic
region and 0.26 (p< 0.01) for the shelf region.
In addition to resolving the dominant spatiotemporal vari-
ability, the model also captures the amplitude of the sea-
sonal surface Chl signal reasonably well. In the oligotrophic
region, the model accurately estimates the observed annual
surface Chl minimum (model: 0.065±0.005 mg Chl m3vs.
SeaWiFS: 0.065 ±0.007 mg Chl m3) while slightly under-
estimating the observed annual maximum (model: 0.47 ±
0.15 mg Chl m3vs. SeaWiFS: 0.75 ±0.55 mg Chl m3).
When model estimates for the entire oligotrophic region are
taken into account (i.e., not restricted to satellite measure-
ment locations and times), the annual minimum develops
in early September, and the maximum develops in late Jan-
uary (Table 1). In the shelf region, greater model–data mis-
match exists for surface Chl, with the model overestimat-
ing the observed annual minimum by 15 % (model: 0.23 ±
0.09 mg Chl m3vs. SeaWiFS: 0.20±0.07 mg Chl m3) and
the observed annual maximum by 102 % (model: 8.09 ±
1.31 mg Chl m3vs. SeaWiFS: 4.01 ±1.23 mg Chl m3).
Here, we find the annual surface Chl seasonal cycle almost
completely out of phase with the oligotrophic region, with
the annual minimum developing in early February and the
annual maximum developing at the end of July (Table 1).
3.2 Regional mesozooplankton biomass model–data
Model mesozooplankton biomass (i.e., LZ +PZ) fields also
agree closely with observations in both the oligotrophic and
shelf region (Fig. 3). Spatial covariance between SEAMAP
climatology and model is statistically significant (p< 0.01),
with a ρvalue of 0.90. When model estimates are com-
pared to SEAMAP tows at corresponding sample times and
locations for the 6835 measurements in the simulation pe-
riod, the ρvalue is 0.55 (p< 0.01). In the oligotrophic
region, the model slightly overestimates mesozooplankton
biomass (model: 4.09±1.82 mg C m3vs. SEAMAP: 3.52±
3.44 mg C m3), with a ρvalue of 0.23 (p< 0.01) and a bias
of 0.57 mg C m3, equivalent to 16 % of the observed mean.
Conversely, in the shelf region, the model underestimates
mesozooplankton biomass (model: 17.40 ±13.58 mg C m3
vs. SEAMAP: 20.91 ±24.62 mg C m3), with a ρvalue of
0.49 (p< 0.01) and a bias of 3.5 mg C m3, equivalent to
17 % of the observed mean. Model estimates and SEAMAP
measurements also compare well with total mesozooplank-
ton biomass measurements (0.2–5 mm) collected in the olig-
otrophic region during the process study cruises (model:
5.55 ±2.87 mg C m3vs. cruise: 4.33 ±2.28 mg C m3).
Although seasonal cycles in the oligotrophic and shelf re-
gions could not be derived from the SEAMAP dataset given
the significant differences in sampling locations over the
course of a year, we investigated model–data mismatches for
each month. The model closely matches or slightly underes-
Biogeosciences, 17, 3385–3407, 2020
T. A. Shropshire et al.: Quantifying spatiotemporal variability in zooplankton dynamics 3393
Figure 2. (a–f) Comparison of surface chlorophyll (mg m3) between SeaWiFS observations (a, c) and model (b, d) from 4 September 1997
to 10 December 2010. Surface chlorophyll climatology computed from daily SeaWiFS images (a) and the corresponding log10 field (c).
Surface chlorophyll climatology computed from daily averages over the entire model simulation (1993–2012) (b) and the corresponding
log10 field (d). Time series of simulated 30 d average surface chlorophyll based on SeaWiFS observations (black) and model (red) for the
oligotrophic region (grid cells with bottom depths 1000 m) (e) and shelf region (grid cells with bottom depths < 1000 m) (f).
timates mesozooplankton biomass for most of the year, with
the exception of January, May and August (Fig. 3a). The
largest model–data mismatch occurs during March, June,
July and December, when the model underestimates meso-
zooplankton biomass by approximately 35 %. Unlike sur-
face Chl, the total mesozooplankton biomass (i.e., depth-
integrated) seasonality is similar in both regions of the GoM.
In the oligotrophic region, the annual biomass minimum
(maximum) occurs at the beginning of January (middle of
May), while in the shelf region, the annual minimum (maxi-
mum) occurs in late December (end of May) (Table 1).
3.3 Chlorophyll and nitrate profile model–data
To validate the vertical structure of the simulated ecosys-
tem, we utilized observed profiles of fluorescence, Chl and
nitrate. When simulated DCM depths were compared to all
2435 SEAMAP fluorescence profiles, we found a statistically
significant correlation (ρ=0.59, p< 0.01) with the observed
maximum fluorescence depth. The maximum fluorescence
depth ranged from the surface to 143 m, while model values
show a similar variability ranging from the surface to 163m
(Fig. 4a). In the oligotrophic region, the model overestimates
the DCM depth (model: 95 ±20 m vs. SEAMAP: 80±25 m)
and has a ρvalue of 0.38 (p< 0.01) with a bias of 15 m,
equivalent to 19% of the observed mean. In the shelf region,
the model also overestimates DCM depth (model: 63 ±26 m
vs. SEAMAP: 53±23 m) and has a ρvalue of 0.49 (p< 0.01)
with a bias of 10 m, equivalent to 19 % of the observed mean.
In contrast, the model slightly underestimated the DCM
depth when compared to calibrated fluorescence profiles
collected during the process cruises (model: 100 ±18 m
vs. observed: 107 ±21 m) (Fig. 4b). In terms of magni-
tude, the model overestimates DCM Chl (model: 0.74 ±
0.35 mg Chl m3vs. observed: 0.38 ±0.13 mg Chl m3), al-
though most of the observations fall within 1 standard de-
viation of the model average. Despite this model–data mis-
match, simulated nitrate profiles closely match profiles from Biogeosciences, 17, 3385–3407, 2020
3394 T. A. Shropshire et al.: Quantifying spatiotemporal variability in zooplankton dynamics
Figure 3. (a–e) Comparison of climatological depth-averaged (0–200m) mesozooplankton biomass (MZB, mg C m3) between SEAMAP
observations (b, d) and model (c, e). Monthly average MZB in SEAMAP observations (black) and model (red) (a) (monthly variability is
not representative of MZB seasonality because sampling locations change between months). MZB climatology computed from SEAMAP
plankton tows (b) and the corresponding log10 field (d). MZB climatology computed from daily averages over the entire model simulation
(1993–2012) (c) and the corresponding log10 field (e).
the World Ocean Database (WOD). In both the model
and observations, the mean nitracline occurred at approx-
imately 75 m (Fig. 4c). On average, model nitrate tended
to be lower at the surface and higher at depth relative
to observations. Above the nitracline, model nitrate was
0.071 ±0.39 mmol N m3while observed nitrate was 0.55 ±
1.29 mmol N m3. Below 200 m, model and data show bet-
ter agreement, with deep nitrate in the model of 24.92 ±3.28
compared to 23.55 ±5.21 mmol N m3in WOD profiles.
DCM depth was evaluated using uncalibrated fluorescence
profiles obtained during SEAMAP cruises. Chlorophyll pro-
files were collected during the May 2017 and 2018 La-
grangian process cruises. For comparisons, the model and
data were sampled at corresponding locations and time of the
year for all simulated years. Nitrate values from the World
Ocean Database that overlapped with the simulation period
and were located in the oligotrophic GoM (> 1000 m) were
used for model–data comparisons.
3.4 Size fractionated mesozooplankton biomass and
grazing model–data comparisons
To further constrain the phytoplankton and zooplankton
community simulated by NEMURO-GoM, we utilized in
situ measurements collected during the process study cruises.
First, we compared the relative proportions of LZ and PZ
biomass to four discrete size classes measured at sea (Fig. 5a,
c). In both the model and observations, we find nearly
identical size distributions assuming that LZ approximates
the smallest two size classes of mesozooplankton sampled
(small mesozooplankton, 0.2–1.0 mm) and PZ approximates
the largest two size classes (large mesozooplankton, 1.0–
5.0 mm). In the field data, small mesozooplankton biomass
varied from 33 % to 46 % (median =40 %, at 95 % C.I.),
while model estimates of LZ biomass vary from 31 % to
46 % (median =40 %). Large mesozooplankton biomass in
the field data varied from 54 % to 67 % (median =60 %),
while model estimates of PZ biomass vary from 54 % to 69%
(median =60 %).
Mesozooplankton specific grazing rates measured during
the process study cruises were also used to validate the sim-
ulated mesozooplankton community. Field measurements
showed that specific grazing rates (µg Chl mg C1d1) de-
creased consistently with increasing mesozooplankton size-
class (Fig. 5b). For model–data comparisons, we com-
puted grazing on LP by LZ and PZ at each depth. Graz-
ing terms were converted into units of Chl using the
model-estimated C :Chl ratio for LP before being depth-
integrated to the corresponding net tow depth and normal-
ized to simulated depth-integrated LZ and PZ biomasses.
We find that model mesozooplankton grazing estimates
Biogeosciences, 17, 3385–3407, 2020
T. A. Shropshire et al.: Quantifying spatiotemporal variability in zooplankton dynamics 3395
Figure 4. (a–c) Model–data comparisons in the vertical distribution of chlorophyll (a, b) and nitrate (c). Comparison between maximum
fluorescence depths observed in SEAMAP profiles and simulated maximum chlorophyll depths (m) (a). Comparison between in situ chloro-
phyll samples collected during May 2017 and 2018 process study cruises and simulated chlorophyll concentrations (mg m3)(b). Gray dots
represent cruise sample values and the red dashed line denotes 1 standard deviation away from the model average. Comparison between
nitrate profiles obtained from the World Ocean Database (WOD) and simulated nitrate (mmol N m3)(c). Gray dots represent WOD sample
values and the red dashed line denotes 1 standard deviation away from the model average.
capture the general trend of decreased specific grazing
rates with increasing mesozooplankton size (Fig. 5d). How-
ever, the model overestimates grazing by small mesozoo-
plankton while underestimating grazing by large mesozoo-
plankton. In the field data, small mesozooplankton graz-
ing ranges from 1.34 to 2.51 µg Chl mg C1d1(median
=1.85) while model estimates of LZ grazing rates vary
from 3.64 to 8.14 µg Chl mg C1d1(median =6.01). Field
measurements of large mesozooplankton grazing range
from 0.76 to 1.44 µg Chl mg C1d1(median =0.94),
while model estimates of PZ grazing vary from 0.44 to
0.70 µg Chl mg C1d1(median =0.58). In terms of to-
tal grazing, the model average is also considerably higher
(2.99 ±2.20 µg Chl mg C1d1) relative to field measure-
ments (1.38 ±0.59 µg Chl mg C1d1) (see Discussion).
3.5 Phytoplankton growth and microzooplankton
grazing model–data comparisons
Measurements of specific phytoplankton growth rates, phy-
toplankton mortality due to microzooplankton grazing and
net primary production (NPP) were used to evaluate pro-
tistan dynamics in the model. We find the model underes-
timates phytoplankton growth and microzooplankton graz-
ing while overestimating NPP (Fig. 6a, b). Phytoplankton
specific growth rates from dilution experiments range from
0.50 to 0.66 d1(median =0.55 d1), while model estimates
of phytoplankton (SP+LP) specific growth rates vary from
0.13 to 0.27 d1(median =0.21 d1). In terms of microzoo-
plankton grazing rates, field data range from 0.19 to 0.55 d1
(median =0.39 d1), while model estimates of SZ vary
from 0.10 to 0.21 d1(median =0.16 d1). NPP estimates
show better agreement with observations, with rates from
276 to 360 mg C m2d1(median =321 mg C m2d1)
in field data, while model estimates vary from 190 to
741 mg C m2d1(median =431 mg C m2d1).
Although the model underestimates phytoplankton growth
and microzooplankton grazing rates, the relative proportion
of NPP consumed by protists in the model (67 %–85 %; me-
dian =76 %) compares reasonably well to field measure-
ments (55 %–92 %; median =72 %) (Fig. 6c). Notably, the
model average proportion of phytoplankton production con-
sumed by protists closely matches the mean for all tropical
waters reported by Calbet and Landry (2004). When phy-
toplankton mortality due to mesozooplankton grazing was
evaluated in the model at cruise sample locations, we found
mesozooplankton grazing accounts for 13 ±8 %, which also
closely agrees with the global average (Calbet et al., 2001).
3.6 Simulated mesozooplankton diet
After model tuning and validation, we utilized NEMURO-
GoM to investigate spatiotemporal variability in diet and sec-
ondary production of the GoM mesozooplankton commu-
nity. First, we examined the trophic level of LZ and PZ in the
model, which provides a measure of their cumulative diet.
Trophic level is calculated by computing the dietary contri-
butions of each prey in LZ (i.e., LP and SZ) and PZ diets
(i.e., LP, SZ and LZ), assuming that the trophic level of LP
equals 1 and that of SZ equals 2. In the oligotrophic region,
both LP and SZ contribute approximately 50 % to LZ diet, as
indicated by a mean trophic level near 2.5 (2.54 ±0.02) for
LZ (Fig. 7a). In the same region, PZ have a trophic level of
2.78 ±0.04, indicating a higher contribution of zooplankton Biogeosciences, 17, 3385–3407, 2020
3396 T. A. Shropshire et al.: Quantifying spatiotemporal variability in zooplankton dynamics
Figure 5. (a–d) Mesozooplankton size-fractioned biomass (a, c) and grazing rates (b, d) in field measurements (black) and model estimates
(red). Mesozooplankton size-fractioned biomass measured during May 2017 and May 2018 process study cruises (a) along with correspond-
ing measured grazing rates (b). Comparisons of model-estimated mesozooplankton biomass to aggregated (size class 1–2 and 3–4) field
measurements (c). Comparisons of model-estimated specific mesozooplankton grazing rates to aggregated (size class 1–2 and 3–4) field
measurements (d). Whiskers extend to 95 % confidence interval. Outliers for model estimates are not shown.
Figure 6. (a–c) Specific phytoplankton growth (µ, d1) and microzooplankton grazing (m, d1) between model (red) and field data
(black) (a). The fraction of phytoplankton growth that is grazed by protists in the model and field data (b). Depth-integrated net primary
production (mg C m2d1)(c). Whiskers extend to the 95 % confidence intervals. Outliers for model estimates are not shown.
to their diet (i.e., SZ and/or LZ) (Fig. 7b). In the shelf re-
gion, LZ are more herbivorous, as indicated by a decrease in
trophic level to 2.31 ±0.01, while PZ are more carnivorous,
as indicated by an increase in trophic level to 2.90 ±0.04.
Despite little evidence for LZ diets dominated by zoo-
plankton in the annual average (in contrast to PZ, which often
have a trophic level 3), we commonly find regions in in-
stantaneous fields during both winter and summer conditions
where SZ are the dominant prey source for LZ (Fig. 7c, e).
These regions, typically in the Loop Current or Loop Cur-
rent eddies (LCEs), highlight the episodic importance of het-
erotrophic protists as prey sources for small mesozooplank-
ton in the GoM. High proportions of SZ in LZ diets can be
attributed to the competitive advantage of SP over LP in ex-
tremely low nutrient environments such as in the Loop Cur-
rent, resulting in high abundances of SP and their predators
(SZ) relative to LP. Instantaneous fields also reveal that phy-
toplankton can be important prey for PZ as well, particularly
during summer, as indicated by trophic levels of around 2.5
in the western GoM (Fig. 7f). In addition to strong variability
Biogeosciences, 17, 3385–3407, 2020
T. A. Shropshire et al.: Quantifying spatiotemporal variability in zooplankton dynamics 3397
in trophic positions, there are also regions in the oligotrophic
GoM, most clearly in the centers of LCEs during summer,
where the model predicts no feeding by mesozooplankton
(Fig. 8e). The convergent anticyclonic circulation of LCEs is
typically associated with low phytoplankton biomass, which
at times may fall near or below feeding thresholds in the NE-
MURO grazing formulation. This formulation is intended to
simulate suppression of feeding activity for zooplankton at
mean prey densities that cannot support the energy expended
while searching for prey.
To investigate which prey contributes most to LZ and PZ
diets, we computed each prey source term for both LZ and
PZ at each grid cell (Fig. 8). As one would expect, the domi-
nant prey for LZ and PZ align closely with spatial variability
in their respective trophic positions. For LZ diet, herbivory
dominates throughout the GoM, except for the Loop Cur-
rent (Fig. 8a). LP contribution to LZ diet is highest on the
shelf, where LP biomass is also high due to the competi-
tive advantage of LP over SP in high-nutrient conditions. In
contrast, PZ diet varies with the relative availability of SZ
and LZ prey. In the oligotrophic region, PZ feed mainly on
SZ (heterotrophic protists) because LZ biomass is relatively
low. On the shelf, they consume primarily LZ (Fig. 8d). De-
spite the significant change in diet between the shelf and
oligotrophic regions, PZ trophic positions remain fairly con-
sistent (Fig. 7d). This is due to the fact that SZ in the olig-
otrophic region and LZ in the shelf region both feed predom-
inantly on phytoplankton and hence occupy similar trophic
levels. In the instantaneous fields for winter (Fig. 8b, e) and
summer (Fig. 8c, f), the dominant prey for both LZ and
PZ show substantial mesoscale variability, indicating that
oceanographic features such as fronts and eddies influence
not only biomass but also zooplankton ecological roles (see
3.7 Simulated mesozooplankton secondary production
To our knowledge, regional secondary production for the
GoM has not been quantified previously. Based on our
model, secondary production due to mesozooplankton aver-
ages 66 ±8×109kg C yr1and ranges from a minimum of
51 ×109kg C (in 1999) to a maximum of 82 ×109kg C (in
2011). In the oligotrophic region, LZ secondary production
averages 35 ±5 mg C m2d1while PZ secondary produc-
tion is 11 ±2 mg C m2d1(Fig. 9). The annual secondary
production minimum develops at the end of December, while
the annual maximum occurs at the beginning of June (Ta-
ble 1). In this region, mesozooplankton secondary production
is 14 ±2×109kg C yr1, equivalent to 6 % of NPP. On the
shelf, secondary production is about 4-fold higher, with LZ
production of 146 ±17 mg C m2d1and PZ production of
42±5 mg C m2d1. Here, the annual minimum also occurs
at the end of December while the maximum occurs later at
the end of July (Table 1). On the shelf, secondary production
constitutes a higher proportion of NPP (13 %) and averages
51 ±6×109kg C yr1.
In addition to differences in total secondary production,
significant differences were found in the mesozooplank-
ton community response to changes in total phytoplankton
biomass on the shelf and in the oligotrophic region. On the
shelf, the average ratio between LZ and PZ secondary pro-
duction is 3.51 and remains almost constant with increasing
phytoplankton biomass (ρ=0.13, p< 0.01). Although we
find a similar average value in the oligotrophic region (3.14),
ratios are more variable and strongly dependent on phyto-
plankton biomass (ρ=0.52, p< 0.01). Ratios of LZ to PZ
secondary production reached values of 2.5 in the lowest
phytoplankton biomass regions of the open-ocean GoM and
increased to 4.0 during times and places where local phy-
toplankton biomass was high. These differences likely stem
from the longer turnover times of PZ, which make them less
sensitive to variability in bottom-up drivers and allow them
to have a proportionally greater role in oligotrophic settings.
As witnessed in the instantaneous fields of diet and sec-
ondary production, mesoscale eddies are common features
in the GoM and hence important to quantify for regional
zooplankton dynamics. To investigate secondary produc-
tion inside cyclonic and anticyclonic eddies, we imple-
ment the TOEddies eddy detection algorithm (Laxenaire et
al., 2018), which uses surface velocities along closed con-
tours of sea surface height (SSH) for detection of mesoscale
eddies (Chaigneau et al., 2011; Laxenaire et al., 2019;
Pegliasco et al., 2015). Grid cells located inside each eddy
are defined within the SSH contour associated with the max-
imum mean surface velocity (interior grid cells). Grid cells
located between the outermost closed contour and within 1.5
radius of the eddy center and not within another eddy were
used to define background conditions outside of eddies (exte-
rior grid cells). Only eddies with areas larger than an equiv-
alent circular diameter of 100 km and not within the Loop
Current were considered in the analysis. On average, 3.78
cyclonic and 3.33 anticyclonic eddies were identified in each
daily velocity field.
In the model, cyclonic eddies were associated with 10 %
higher secondary production relative to exterior grid cells,
and the ratio of secondary production in interior cells to ex-
terior cells ranged from 0.4 to 3.37 (95 % CI). In contrast,
secondary production was substantially lower inside anticy-
clonic eddies, accounting for only 46 % of the average sec-
ondary production in exterior cells (0.03–1.87, 95 % CI). In
addition to their convergent nature that dampens nutrient in-
put, lower rates of secondary production in anticyclonic ed-
dies can likely be attributed to the presence of highly olig-
otrophic Loop Current water trapped within large anticy-
clonic LCEs. Biogeosciences, 17, 3385–3407, 2020
3398 T. A. Shropshire et al.: Quantifying spatiotemporal variability in zooplankton dynamics
Figure 7. (a–f) Trophic levels of simulated large zooplankton (LZ, a, b, c) and predatory zooplankton (PZ, d, e, f). Annual-averaged
trophic positions of LZ (a) and PZ (d). Instantaneous model field of trophic positions during winter conditions for LZ (b) and PZ (e) on
4 February 2012. Instantaneous model field of trophic positions during summer conditions for LZ (c) and PZ (f) on 2 August 2011.
Figure 8. (a–f) Dominant prey source for simulated large zooplankton (LZ, a, b, c) and predatory zooplankton (PZ, d, e, f). Annual-averaged
diet contributions for LZ (a) and PZ (d). Instantaneous model field of zooplankton diet during winter conditions for LZ (b) and PZ (e) on
4 February 2012. Instantaneous model field of zooplankton diet during summer conditions for LZ (c) and PZ (f) on 2 August 2011.
4 Discussion
Many parameters in biogeochemical models are poorly con-
strained by observations and laboratory studies and/or highly
variable in the environment. The numbers and uncertainties
around these parameters allow PBMs with varying degrees
of tuning to reproduce a single ecosystem attribute (e.g., sur-
face Chl) even if multiple processes are inaccurately repre-
sented (Anderson, 2005; Franks, 2009). Once validated, one
of the main values of coupling physical and biogeochemi-
cal models (i.e., PBMs) is their utility for making inferences
about portions of the lower trophic level that are undersam-
pled and/or difficult to measure in the field. If PBMs are to be
utilized to explain variability rather than simply fit an obser-
vational dataset, multiple ecosystem attributes must be val-
idated and the underlying model structure and assumptions
critically evaluated. In the section below, we further justify
changes to model structure by evaluating the underlying as-
sumptions in default NEMURO, and we discuss model–data
mismatch before drawing conclusions about the GoM zoo-
plankton community and the implications of its dynamics for
higher trophic levels.
4.1 Justification for NEMURO modifications
The phytoplankton community in the North Pacific (NP) do-
main where NEMURO was originally designed is largely
composed of nanoplankton (original SP) and microplankton
(original LP). By default, SP is assumed to represent coccol-
ithophores and autotrophic nanoflagellates, which can be im-
portant prey of copepods and other mesozooplankton in tem-
perate and subpolar regions (Kishi et al., 2007). However, in
tropical regions such as the GoM, smaller picophytoplankton
taxa typically dominate, particularly in highly oligotrophic
Biogeosciences, 17, 3385–3407, 2020
T. A. Shropshire et al.: Quantifying spatiotemporal variability in zooplankton dynamics 3399
Figure 9. (a–f) Vertically integrated secondary production (mg C m2d1) of simulated large zooplankton (LZ, a, b, c) and predatory
zooplankton (PZ, d, e, f). Annual-averaged secondary production for LZ (a) and PZ (d). Instantaneous model field of secondary production
during winter conditions for LZ (b) and PZ (e) on 4 February 2012. Instantaneous model field of secondary production during summer
conditions for LZ (c) and PZ (f) on 2 August 2011.
Table 1. Average seasonal minimum and maximum values in the model (1993–2012) and the day of year in which they occur for surface
chlorophyll (mg m3), depth-integrated estimates of phytoplankton biomass (mg C m2), net primary production (mg Cm2d1), meso-
zooplankton biomass (mg C m2), and mesozooplankton secondary production (mgC m2d1) calculated by spatially averaging daily fields
over the oligotrophic region (grid cells with bottom depths > 1000 m) and shelf region (grid cells with bottom depths <1000 m). Day of year
values are in the format “month/day ±days”.
Daily field value Day of year
Diagnostic (oligotrophic) Annual min. Annual max. Day of min. Day of max.
Surface chlorophyll 0.09 ±0.005 0.27 ±0.06 9/9±23 1/29 ±13
Phytoplankton biomass 2300 ±130 3600 ±140 12/26 ±7 4/29 ±17
Net primary production 290 ±70 1000 ±120 12/31 ±12 7/6±27
Mesozooplankton biomass 1000 ±40 1400 ±90 1/1±4 5/19 ±18
Secondary production 18 ±4 68 ±10 12/31 ±10 6/4±15
Diagnostic (shelf) Annual min. Annual max. Day of min. Day of max.
Surface chlorophyll 1.96 ±0.15 3.00 ±0.30 2/8±37 7/31 ±58
Phytoplankton biomass 3200 ±290 5200 ±440 1/1±9 7/18 ±11
Net primary production 750 ±120 2000±220 12/31 ±8 7/21 ±14
Mesozooplankton biomass 670 ±70 1100 ±90 12/29 ±7 5/23 ±25
Secondary production 94 ±17 270 ±28 12/31 ±6 7/20 ±16
regions. Common picophytoplankton of the GoM include
cyanobacteria and picoeukaryotes, which are too small for
most mesozooplankton to feed on. Hence, the SP-to-LZ graz-
ing pathway was removed in our model. We found that re-
moval of this grazing pathway allowed the model to simulate
a more realistic phytoplankton community on the shelf re-
gion. Despite intuition, SP largely dominated the shelf region
in the model when LZ were allowed to graze on SP. After
closer inspection, we found that grazing of SP sustained LZ
biomass on the shelf to levels where top-down pressure con-
strained LP standing stocks. This prevented large blooms of
LP, leading to a competitive advantage for SP even in highly
eutrophic conditions (the Mississippi River delta), which was
observed for a wide range of LP maximum growth rates, LP
half-saturation constants, and LZ and PZ grazing rates. Thus,
removal of the SP to LZ grazing pathway added ecological
realism and improved the model solution.
During the model tuning process (outlined in the Supple-
ment), we also found that, despite a wide range of tested
parameter sets, the model was unable to simulate mesozoo-
plankton biomass low enough to match SEAMAP observa-
tions in the oligotrophic region. Even with unrealistically low
phytoplankton biomass, equivalent to approximately 50% of
surface Chl observed in SeaWiFS images, the model overes-
timated mesozooplankton biomass. To achieve realistic lev-
els of mesozooplankton biomass in the oligotrophic region, Biogeosciences, 17, 3385–3407, 2020
3400 T. A. Shropshire et al.: Quantifying spatiotemporal variability in zooplankton dynamics
default LZ and PZ mortality parameter values needed to be
increased by an order of magnitude. However, this produced
unrealistically high loss rates on the shelf region, leading to
mesozooplankton biomass estimates that were substantially
lower than SEAMAP shelf observations. Implementation of
linear mortality on all biological state variables (except PZ)
resolved this issue by providing the model with a greater dy-
namic range. In NEMURO, and other biogeochemical mod-
els, quadratic mortality is often used to increase model stabil-
ity and/or is mechanistically justified as representing the im-
pact of unmodeled predators that covary in abundance with
prey (Gentleman and Neuheimer, 2008; Steele and Hender-
son, 1992). However, grazing losses of all state variables (ex-
cept PZ) are already explicitly modeled in NEMURO by de-
fault. Hence, removal of quadratic mortality also added eco-
logical realism and improved the model solution. Quadratic
mortality was retained for PZ to account for the implicit pre-
dation pressure of unmodeled predators (e.g., planktivorous
4.2 Model–data mismatch
4.2.1 Surface chlorophyll discrepancies
Within our model–data comparisons of surface Chl we find
that NEMURO-GoM reproduces important patterns in both
the oligotrophic and shelf region, the latter of which, apart
from the northern shelf, has not been well resolved by pre-
vious PBMs (e.g., Gomez et al., 2018; Xue et al., 2013).
The absence of a shelf Chl signature may, in some cases,
be overly attributed to bias in satellite measurement due
to high concentrations of colored dissolved organic matter
(CDOM). While a clear shelf signature is well resolved in
NEMURO-GoM, the model–data mismatch is greater on the
shelf compared to oligotrophic regions. This is an expected
result considering that the model incorporates climatological
river forcing while actual variability is much more complex.
Furthermore, the absence of CDOM in the model likely con-
tributes to the overestimation of phytoplankton biomass on
the shelf.
In future studies, the inclusion of daily nutrient data like
that produced for the Mississippi River by USGS starting
in 2011 is needed for PBMs to better resolve variability on
the shelf. Including benthic processes, such as denitrification
(Fennel et al., 2006), may also reduce model–data mismatch
in shelf regions. Implementing more realistic light attenua-
tion (e.g., wavelength-specific light attenuation or inclusion
of CDOM) could further improve estimates of phytoplankton
biomass on the shelf as primary production can be sensitive
to different light attenuation formulations (Anderson et al.,
2015). In our model, it was difficult to simulate deep DCMs
in the oligotrophic GoM while also simulating DCMs on the
shelf that were shallow enough to maintain high nitrate. This
may reflect the need for more realistic light attenuation in the
Quantifying uncertainty in C :Chl ratios is also an impor-
tant task moving forward as future PBMs will likely con-
tinue to depend heavily on satellite Chl for the bulk of model
validation. To correctly parameterize C :Chl ratios, more in
situ samples are needed that resolve changes in phytoplank-
ton light-harvesting pigments along gradients from coastal to
oligotrophic regions and from the surface to the DCM. With-
out these observations, it is difficult to gauge mismatches
between model and satellite ocean color products (e.g., dis-
crepancies on the Campeche Bank in our model) or in situ
profiles of Chl. In addition to validation, these measurements
are needed to avoid erroneous model tuning. For instance, a
model that exhibits significant mismatch with respect to sur-
face Chl may in fact accurately estimate carbon-based phyto-
plankton biomass while using unrealistic C :Chl ratios. One
could arrive at incorrect conclusions about regional ecosys-
tem dynamics as a result of modifying model parameters or
structure in an effort to better fit Chl observations. Given the
importance of C :Chl ratios in PBMs, future studies should
quantify uncertainty in modeled Chl through sensitivity ex-
periments focused on C :Chl model parameters and formu-
lation, with explicit comparison to direct field measurements
of phytoplankton C :Chl.
In our model, the most noticeable surface Chl model–
data mismatch occurs on the southern GoM shelf (Campeche
Bank), where the model consistently overestimates surface
Chl. This bias was also notable in the GoM PBM imple-
mented by Damien et al. (2018), particularly in winter. We
believe this discrepancy is driven by a combination of er-
rors involving overestimation of shelf mixing by the hydro-
dynamic model, entrainment of high-Chl water (given the
overestimated DCM magnitude in the model) or errors in the
open boundary conditions which result in an overestimation
of upwelled nutrients/biomass near the Yucatán Peninsula
that are transported westward by shelf currents. We found
that the Campeche Bank model–data mismatch was reduced
when open boundary conditions included nitracline depths
of greater than 100 m. This may reflect realistic in situ condi-
tions considering that Caribbean water entering the GoM is
highly oligotrophic. During our process cruises, nitrate was
often undetectable above 100 m in samples collected near
the Loop Current (Angela Knapp, personal communication,
Although modifying the boundary conditions may be jus-
tified, deepening the nitracline at the boundaries made it in-
creasingly difficult to sustain realistic surface phytoplankton
biomass in the oligotrophic GoM. This may point to the im-
portance of nitrogen-fixing cyanobacteria, which provide an
alternative source of new nitrogen (other than upwelling and
mixing) that could be supporting phytoplankton at the sur-
face given the strong stratification and deep nitraclines in the
GoM. In the process of model tuning, we noticed that in-
creasing the DON pool by increasing the PON to DON de-
composition rate was necessary to maintain both relatively
deep nitraclines and realistic surface Chl by providing a slow
Biogeosciences, 17, 3385–3407, 2020
T. A. Shropshire et al.: Quantifying spatiotemporal variability in zooplankton dynamics 3401
leeching of ammonium near the surface through bacterial
communities. The need for this slow production of ammo-
nium in surface layers may compensate for nitrogen fixation,
which is not included in NEMURO (Holl et al., 2007; Mul-
holland et al., 2006). In future studies, including diazotrophs
as a separate phytoplankton functional type would be essen-
tial for evaluating the importance of nitrogen fixation in the
Despite the model–data mismatch on the Campeche Bank,
this discrepancy appears to have little impact on the rest of
the GoM. However, the model overestimates surface Chl in
the southwestern GoM, which can likely be attributed to en-
trainment of high-Chl water originating from the Campeche
Bank. Locally, the ecological impact is likely more signif-
icant. Higher phytoplankton biomass would be expected to
support higher mesozooplankton grazing rates and secondary
production. Indeed, some of the highest model rates of sec-
ondary production occur on the Campeche Bank. Hence, the
surface Chl model–data mismatch may lead to an overesti-
mation of secondary production for this region.
4.2.2 Deep chlorophyll maximum discrepancies
Since most PBMs focus on validating against satellite-
derived surface chlorophyll, the dynamics of the DCM is
often insufficiently investigated. Consequently, many mod-
els predict DCM depths that are far too shallow. Identifying
this issue in the literature proved to be difficult because most
studies do not provide profiles of simulated Chl (an exception
is the recent GoM PBM by Damien et al., 2018). We note that
DCM depths in the DIAZO model (Stukel et al., 2014) were
often quite shallow or completely nonexistent in the portion
of the domain that included the oligotrophic GoM region.
Underestimates of DCM depth in the unmodified COBALT
biogeochemical model have also been identified (Moeller
et al., 2019). In our investigation of the GoM PBM imple-
mented by Gomez et al. (2018), we found that DCMs in the
oligotrophic region were commonly shallow and weak. In
the default NEMURO simulation, DCM depths in the olig-
otrophic region were typically at a depth of 25 m, which is
much shallower than observed (SEAMAP: 80 ±25 m, Pro-
cess cruises: 107±21 m). While this issue may seem insignif-
icant, particularly if a study is focused on mixed-layer dy-
namics, accurate placement of the DCM can have profound
impacts on PBM behaviors, because the DCM is typically
colocated with the nitracline. Unrealistically shallow DCMs
and nitraclines permit unrealistically high nitrate fluxes into
the surface layer following mixing events; thus, validating
the DCM in PBMs is critical.
For these reasons, we devoted substantial effort to tuning
phytoplankton dynamics in the DCM. Modifications to α(the
slope of the photosynthesis–irradiance curve) and attenua-
tion coefficients allowed the model to estimate deeper more
realistic DCM depths. Inclusion of a variable C :Chl module
was also implemented to better resolve the DCM. However,
an additional issue was present in the default NEMURO sim-
ulations, the NEMURO-GoM, and every simulation that we
attempted. In all simulations that formed DCMs, the location
of the DCM was always colocated with a maximum in phy-
toplankton specific growth rate, even though field measure-
ments indicate that phytoplankton growth rates and NPP are
either relatively constant with depth or decline in the DCM.
This is not surprising, given the low photon flux at the base
of the euphotic zone and the energetic demands required to
upregulate cellular density of light-harvesting pigments. Ad-
ditionally, our field measurements show that the DCM was
not associated with a biomass maximum (biomass was fairly
constant with depth), suggesting that DCM formation in the
GoM is physiologically driven.
We believe this DCM dynamical issue was responsible,
in part, for the underestimation of specific phytoplankton
growth and microzooplankton grazing rates by the model de-
spite estimating higher NPP (Fig. 4d). An underestimation
of C :Chl ratios or overestimation of phytoplankton biomass
may also contribute to the model–data mismatch at the DCM.
The ecological impact of this mismatch is likely to be great-
est in the model during the summer months when the mixed-
layer depth is shallow and the DCM reaches its maximum
depth. Elevated growth rates or biomass at the DCM during
this period could significantly decrease nitrate flux into the
mixed layer, causing the model to underestimate near-surface
primary and secondary production.
Future PBM studies need to focus more effort on resolv-
ing ecological dynamics responsible for the formation of the
DCM. Errors originating from the hydrodynamic model or
use of temporally averaged velocity fields used in offline
models may also contribute to model–data mismatch at the
DCM. Vertical mixing is particularly important to PBMs but
is often poorly validated in hydrodynamic models. Greater
coordination between physical modelers and biologists to
constrain vertical fluxes should be considered an important
avenue for improving PBM simulations moving forward.
4.2.3 Mesozooplankton grazing discrepancies
Novel to this study, model estimates of mesozooplankton
biomass are shown to agree closely with observations on the
shelf and in the oligotrophic GoM. To our knowledge, this
study includes the first quasi-regional zooplankton biomass
model validation in a PBM. Our model also provides the
first model–data comparisons of size-specific zooplankton
biomass and grazing rates for the GoM. Such comparisons
provide valuable insights into the potential biases of tra-
ditional functional-group biogeochemical models pertain-
ing to zooplankton dynamics (Everett et al., 2017). While
NEMURO-GoM shows broad agreement with zooplankton
observations, some model–data mismatch occurs, particu-
larly for mesozooplankton grazing rates.
We identify three factors that may explain the model–data
mismatch for mesozooplankton grazing rates. The first and Biogeosciences, 17, 3385–3407, 2020
3402 T. A. Shropshire et al.: Quantifying spatiotemporal variability in zooplankton dynamics
most obvious factor is the temporal sampling discrepancy
as measurements were collected outside our model simula-
tion period. Model–data mismatch may also arise from in-
accuracies in the field measurements. During our process
cruises, the zooplankton gut pigment measurements were
based solely on phaeopigment content due to phytodetrital
aggregates and Trichodesmium colonies found in our zoo-
plankton net tows, which can lead to substantial contamina-
tion. Thus, true mesozooplankton grazing rates were likely
underestimated because undegraded Chl can be abundant in
the foreguts of mesozooplankton. Furthermore, the gut pig-
ment approach assumes that any group of mesozooplankton
has a constant gut throughput time (as a function of temper-
ature), which is an oversimplification.
Uncertainties in model grazing formulations could also
contribute to model–data mismatch (Gentleman et al., 2003a;
Sailley et al., 2015). Future in situ grazing measurements
are needed to enable an objective selection of grazing for-
mulations and parameter values. In particular, field studies
that shed light on prey selectivity would be useful for pa-
rameterizing PBMs with multiple mesozooplankton func-
tional groups, such as NEMURO-GoM. Such studies are
challenging, however, because the difficulty of making in
situ grazing measurements on mesozooplankton, combined
with the inherent uncertainty of these measurements, can
make it challenging to differentiate between, for instance,
Ivlev and Holling’s disk grazing formulations (e.g., Fig. 4
in Morrow et al., 2018). Nevertheless, differences in param-
eterizing grazing can lead to substantially different model
behavior (Anderson et al., 2010; Sailley et al., 2015; Wain-
wright et al., 2007). In NEMURO-GoM, secondary produc-
tion and dietary preferences of the mesozooplankton com-
munity are both strongly influenced by model grazing for-
mulation. While we carefully chose parameterizations that
gave reasonable fits to extensive field datasets of zooplank-
ton biomass and grazing rates, this does not preclude the pos-
sibility that other functional forms would have more accu-
rately simulated zooplankton dynamics. Hence, future PBMs
should investigate how different grazing formulations impact
zooplankton dynamics in the region. We especially recom-
mend collaborations between experimentalists (potentially
using new techniques such as DNA metabarcoding of gut
contents) and modelers to develop synthetic approaches with
the potential to quantitatively assess the realism of different
grazing formulations.
Clear model–data mismatch is also evident in the propor-
tion of grazing mediated by PZ and LZ. This may be due to
the fact that PZ is by default explicitly defined and parameter-
ized as a higher-trophic-level mesozooplankton that can feed
on LZ. In reality, while there is a correlation between size
and trophic level in the ocean, many predatory zooplankton
are < 1 mm, and many suspension-feeding zooplankton are
> 1 mm; hence, the overlap of taxonomic groups with differ-
ent functional roles and sizes makes it difficult to directly
compare model categories to field data. For example, shelf
suspension-feeding zooplankton are likely larger than their
counterparts in the oligotrophic GoM although their func-
tional role in the ecosystem does not change between envi-
The ecological impact of the model’s potential overesti-
mation of LZ grazing rates is most likely to manifest through
an increase in the ratio of secondary production to meso-
zooplankton biomass. Since both LZ and PZ biomasses are
accurately modeled by NEMURO-GoM, the overestimation
of grazing rates suggests that LZ turnover times may be too
high, thus leading to higher estimates of secondary produc-
tion. However, this interpretation may oversimplify the com-
plex interactions within pelagic protistan communities. In the
oligotrophic region where our model overestimates LZ graz-
ing rates, the model indicates that heterotrophic protists com-
prise approximately half of the LZ diet. Thus, overestimates
of grazing on LP do not necessarily lead to overestimates in
total consumption if < 1 mm zooplankton derive substantial
nutrition from nonphototrophic sources in the field. Further-
more, the model’s construction (i.e., LZ and PZ are func-
tional groups, while the field data are size classes) suggests
that part of the model–data mismatch in Fig. 5d may re-
sult from the presence of some suspension feeders (i.e., LZ)
in the > 1 mm zooplankton size class and some carnivorous
zooplankton (i.e., PZ) in the < 1 mm zooplankton size class.
In this case, the model may simply attribute too high of a
LP :SZ prey ratio to LZ. If this is the issue, the model’s es-
timate of LZ secondary production may be accurate but its
trophic level too low (or, conversely, the trophic level of PZ
too high). Direct assessments of zooplankton trophic posi-
tion (e.g., by compound-specific isotopic analysis of amino
acids, Chikaraishi et al., 2009; Décima et al., 2017) may help
resolve these issues.
4.3 Mesozooplankton dynamics in the open-ocean
oligotrophic Gulf of Mexico
Despite its nutrient-poor conditions, the open-ocean GoM
ecosystem is a key region for spawning and larval develop-
ment of many commercially important fishes, including At-
lantic bluefin tuna, yellowfin tuna, skipjack tuna, sailfish and
mahi mahi (Cornic and Rooker, 2018; Kitchens and Rooker,
2014; Lindo-Atichati et al., 2012; Muhling et al., 2017;
Rooker et al., 2012, 2013). Why so many species choose
such oligotrophic waters as habitat for their larval stages is
unknown but may be due to reduced predation risk (Bakun,
2013; Bakun and Broad, 2003). Regardless, rapid growth and
survival through the larval period depend on mesozooplank-
ton prey that are suitably abundant and appropriately sized
for these larval fishes. These prey taxa may be especially
sensitive to increased stratification and oligotrophication as-
sociated with climate change, making investigation of their
dynamics and production an important topic of research.
Mesozooplankton biomass in the oligotrophic GoM was
found to be strikingly low in both observations and model es-
Biogeosciences, 17, 3385–3407, 2020
T. A. Shropshire et al.: Quantifying spatiotemporal variability in zooplankton dynamics 3403
timates, approximately an order of magnitude less than on the
shelf. Model results clearly show that this low biomass con-
dition arises from bottom-up resource limitation. Our results
suggest that low phytoplankton biomass in oligotrophic re-
gions, and particularly within Loop Current eddies, may even
lead to localized and episodic regions where phytoplankton
concentrations approach thresholds low enough that trigger
collapse of mesozooplankton grazing. Prey-limiting condi-
tions for mesozooplankton and their predators would be ex-
pected more frequently in the GoM during warmer ocean
conditions. Higher sea surface temperatures and increased
thermal stratification could suppress vertical mixing, result-
ing in lower phytoplankton biomass.
Despite extreme oligotrophy and dominance of picophy-
toplankton, our model shows that both PZ and LZ can be
sustained at modest abundances in the oligotrophic GoM.
Indeed, the substantial abundances of large (> 1 mm) meso-
zooplankton, equivalent to 60% of total mesozooplankton in
both observations and model results (Fig. 4a, c), are an im-
portant result that helps explain the success of larval fish in
the region. Our results show that large mesozooplankton (PZ)
occupy a trophic position of approximately 3.0 in the open-
ocean GoM, which is marginally lower than on the shelf
where they feed primarily on small mesozooplankton (LZ).
The change in trophic position is associated with a switch
from carnivory to feeding predominantly on heterotrophic
protists in the oligotrophic region. This result highlights the
importance of intermediate protistan trophic levels in sustain-
ing mesozooplankton communities in oligotrophic regions.
Indeed, both LZ and PZ ingest proportionally more SZ in
the open ocean than on the shelf. Notably, these protistan
trophic steps cannot be quantified by routine field techniques
because they have no pigment signature to make them visi-
ble in gut pigment measurements and may not enrich in bulk
15N, leading to isotopic invisibility from a trophic perspec-
tive (Gutiérrez-Rodríguez et al., 2014). Despite their impor-
tance to phytoplankton grazing, they are sometimes missing
from GoM ecosystem models (e.g., Fennel et al., 2011) and
severely underrepresented or even absent in complex mass-
balance constrained models (Arreguin-Sanchez et al., 2004;
Geers et al., 2016). New insights may arise from focused in-
vestigations of phytoplankton–protist–crustacean linkages in
oligotrophic regions in both model and experimental stud-
ies. This will likely require the use of next-generation tech-
nologies such as compound-specific isotopic analyses of spe-
cific amino acids that have been shown to enrich in protists
(Décima et al., 2017) or DNA metabarcoding to assess zoo-
plankton gut contents (Cleary et al., 2016).
Another robust finding from this study is the dynamic
mesoscale variability in zooplankton abundance, diet and
trophic position. These model results highlight the impact
of Loop Current eddies and mesoscale fronts and other fea-
tures in modifying the biogeochemistry and food web of the
GoM. The existence of hot spots of productivity in the GoM
has been noted in observational studies (Biggs and Ressler,
2001), and the importance of GoM mesoscale features to
fish larvae has been hypothesized (Domingues et al., 2016;
Lindo-Atichati et al., 2012; Rooker et al., 2012). Indeed, cy-
clonic eddies were found to have enhanced secondary pro-
duction in our model, while secondary production was de-
pressed within anticyclonic eddies. Our results further sug-
gest that these mesoscale structures may not only modify
zooplankton abundances, but also their trophic roles in the
ecosystem, with implications for the transfer efficiencies of
carbon and nitrogen in the pelagic food web.
5 Conclusions
In this study we developed a PBM for the GoM and after ex-
tensive validation used the model to investigate zooplankton
community dynamics. The model was able to capture broad
ecosystem attributes including phytoplankton and mesozoo-
plankton abundances, depth of the DCM and nutricline, and
growth and grazing patterns. Based on model estimates of
trophic position and diet our results suggest that on the shelf
small mesozooplankton are strongly herbivorous while large
mesozooplankton are strongly carnivorous. However, distinct
changes in diet were noted in the oligotrophic GoM, where
both groups rely more on protistan prey. Changes in diet and
secondary production highlighted in this study have the po-
tential to impact food availability to higher trophic levels,
such as pelagic larval fishes. In future work, we plan to cou-
ple NEMURO-GoM to an individual-based model to evalu-
ate the extent to which mesozooplankton abundance limits
larval fish feeding, growth and survival along transport path-
ways in the GoM. Insights from this ecosystem-based ap-
proach may help to better resolve stock-recruitment relation-
ships that are needed for sustainable fisheries management
and improved stock-assessment models.
Code and data availability. The model code and model valida-
tion data used in this study can be downloaded from GitHub at (Shropshire, 2019a).
An idealized one-dimensional version of NEMURO-GoM written
in MATLAB is also provided. The three-dimensional NEMURO-
GoM model outputs used in the study are available on the FSU-
COAPS server in a Network Common Data Form (NetCDF for-
mat). Data are also publicly available through the Gulf of Mex-
ico Research Initiative Information and Data Cooperative (GRI-
IDC) at (Shropshire, 2019b,
Supplement. The supplement related to this article is available on-
line at:
Author contributions. TAS conducted all numerical simulations
and model analysis. EPC, SLM and AB provided expertise on the Biogeosciences, 17, 3385–3407, 2020
3404 T. A. Shropshire et al.: Quantifying spatiotemporal variability in zooplankton dynamics
hydrodynamic modeling. MRS and VJC provided expertise on the
biogeochemical model coding and tuning. RS, MRS, MRL and GZ
processed and provided data that were central to NEMURO-GoM’s
validation. TAS wrote the manuscript with contributions from all
Competing interests. The authors declare that they have no conflict
of interest.
Acknowledgements. We thank the captains and crew of the NOAA
ship Nancy Foster and many of our colleagues from NOAA SEFSC
and the NASA-funded Predicting, Validating, and Understanding
Zooplankton Distributions from Space in an Eddy Rich Ocean
project. We also thank Oliver Jahn for providing valuable direction
in configuring the offline MITgcm package as well as Mandy Kar-
nauskas, Sang-Ki Lee and Fabian Gomez for their input during
model development. We also thank Rémi Laxenaire for assistance
with the TOEddies eddy detection algorithm.
Financial support. This paper is a result of research supported by
the Gulf of Mexico Research Initiative under the CSOMIO project,
the NOAA RESTORE Science Program under federal funding op-
portunity NOAA-NOS-NCCOS-2017-2004875, the NOAA Center
for Coastal and Marine Ecosystems (NOAA Office of Education,
Educational Partnership Program, award NA16SEC48100009), the
NOAA NMFS Fisheries and the Environment program, the North-
ern Gulf Institute (projects 18-NGI3-41 and 18-NGI3-52) under
NOAA award NA16OAR4320199, and by NASA IDS grant no.
Review statement. This paper was edited by Katja Fennel and re-
viewed by three anonymous referees.
Anderson, T. R.: Plankton functional type modelling: Run-
ning before we can walk?, J. Plankton Res., 27, 1073–1081,, 2005.
Anderson, T. R., Gentleman, W. C., and Sinha, B.: In-
fluence of grazing formulations on the emergent proper-
ties of a complex ecosystem model in a global ocean
general circulation model, Prog. Oceanogr., 87, 201–213,, 2010.
Anderson, T. R., Gentleman, W. C., and Yool, A.: EMPOWER-
1.0: an Efficient Model of Planktonic ecOsystems WrittEn in R,
Geosci. Model Dev., 8, 2231–2262,
8-2231-2015, 2015.
Arreguin-Sanchez, F., Zetina-Rejón, M., Manickchand-Heileman,
S., Ramírez-Rodríguez, M., and Vidal, L.: Simulated response
to harvesting strategies in an exploited ecosystem in the
southwestern Gulf of Mexico, Ecol. Model., 172, 421–432,, 2004.
Bakun, A.: Ocean eddies, predator pits and bluefin tuna: Impli-
cations of an inferred “low risk-limited payoff” reproductive
scheme of a (former) archetypical top predator, Fish Fish., 14,
424–438,, 2013.
Bakun, A. and Broad, K.: Environmental “loopholes” and fish pop-
ulation dynamics: Comparative pattern recognition with focus
on El Niño effects in the Pacific, Fish. Oceanogr., 12, 458–473,, 2003.
Biggs, D. C. and Ressler, P. H.: Distribution and abundance of
phytoplankton, zooplankton, icthyoplankton, and micronekton
in the deepwater Gulf of Mexico, Gulf Mex. Sci., 19, 7–29,, 2001.
Buitenhuis, E., Le Quéré, C., Aumont, O., Beaugrand, G.,
Bunker, A., Hirst, A., Ikeda, T., O’Brien, T., Piontkovski,
S., and Straile, D.: Biogeochemical fluxes through meso-
zooplankton, Global Biogeochem. Cy., 20, GB2003,, 2006.
Calbet, A.: Mesozooplankton grazing effect on primary
production: A global comparative analysis in ma-
rine ecosystems, Limnol. Oceanogr., 46, 1824–1830,, 2001.
Calbet, A. and Landry, M. R.: Phytoplankton growth,
microzooplankton grazing, and carbon cycling in
marine systems, Limnol. Oceanogr., 49, 51–57,, 2004.
Caron, D. A. and Hutchins, D. A.: The effects of changing climate
on microzooplankton grazing and community structure: Drivers,
predictions and knowledge gaps, J. Plankton Res., 35, 235–252,, 2013.
Chaigneau, A., Le Texier, M., Eldin, G., Grados, C., and Pizarro,
O.: Vertical structure of mesoscale eddies in the eastern
South Pacific Ocean: A composite analysis from altimetry and
Argo profiling floats, J. Geophys. Res.-Oceans, 116, 1–16,, 2011.
Chassignet, E. P., Smith, L. T., Halliwell, G. R., and Bleck,
R.: North Atlantic simulations with the Hybrid Coordi-
nate Ocean Model (HYCOM): Impact of the vertical co-
ordinate choice, reference pressure, and thermobaricity, J.
Phys. Oceanogr., 33, 2504–2526,
0485(2003)033<2504:NASWTH>2.0.CO;2, 2003.
Chikaraishi, Y., Ogawa, N. O., Kashiyama, Y., Takano, Y.,
Suga, H., Tomitani, A., Miyashita, H., Kitazato, H. and
Ohkouchi, N.: Determination of aquatic food-web structure
based on compound-specific nitrogen isotopic composition
of amino acids, Limnol. Oceanogr.-Methods, 7, 740–750,, 2009.
Cleary, A. C., Durbin, E. G., Rynearson, T. A., and Bailey,
J.: Feeding by Pseudocalanus copepods in the Bering
Sea: Trophic linkages and a potential mechanism of
niche partitioning, Deep-Sea Res. Pt. II, 134, 181–189,, 2016.
Coles, V. J., Stukel, M. R., Brooks, M. T., Burd, A., Crump, B. C.,
Moran, M. A., Paul, J. H., Satinsky, B. M., Yager, P. L., Zielin-
ski, B. L., and Hood, R. R.: Ocean biogeochemistry modeled
with emergent trait-based genomics, Science, 358, 1149–1154,, 2017.
Cornic, M. and Rooker, J. R.: Influence of oceanographic conditions
on the distribution and abundance of blackfin tuna (Thunnus at-
Biogeosciences, 17, 3385–3407, 2020
T. A. Shropshire et al.: Quantifying spatiotemporal variability in zooplankton dynamics 3405
lanticus) larvae in the Gulf of Mexico, Fish. Res., 201, 1–10,, 2018.
Damien, P., Pasqueron de Fommervault, O., Sheinbaum, J.,
Jouanno, J., Camacho-Ibar, V. F., and Duteil, O.: Partition-
ing of the Open Waters of the Gulf of Mexico Based
on the Seasonal and Interannual Variability of Chlorophyll
Concentration, J. Geophys. Res.-Oceans, 123, 2592–2614,, 2018.
Décima, M., Landry, M. R., and Rykaczewski, R. R.: Broad
scale patterns in mesozooplankton biomass and grazing in the
eastern equatorial Pacific, Deep-Sea Res. Pt. II, 58, 387–399,, 2011.
Décima, M., Landry, M. R., Stukel, M. R., Lopez-Lopez, L.,
and Krause, J. W.: Mesozooplankton biomass and grazing
in the Costa Rica Dome: Amplifying variability through
the plankton food web, J. Plankton Res., 38, 317–330,, 2016.
Décima, M., Landry, M. R., Bradley, C. J., and Fogel, M. L.: Ala-
nine δ15N trophic fractionation in heterotrophic protists, Limnol.
Oceanogr., 62, 2308–2322,,
Domingues, R., Goni, G., Bringas, F., Muhling, B., Lindo-Atichati,
D., and Walter, J.: Variability of preferred environmental condi-
tions for Atlantic bluefin tuna (Thunnus thynnus) larvae in the
Gulf of Mexico during 1993–2011, Fish. Oceanogr., 25, 320–
336,, 2016.
Doney, S. C., Lima, I., Moore, J. K., Lindsay, K., Behren-
feld, M. J., Westberry, T. K., Mahowald, N., Glover, D.
M., and Takahashi, T.: Skill metrics for confronting global
upper ocean ecosystem-biogeochemistry models against field
and remote sensing data, J. Marine Syst., 76, 95–112,, 2009.
Everett, J. D., Baird, M. E., Buchanan, P., Bulman, C., Davies, C.,
Downie, R., Griffiths, C., Heneghan, R., Kloser, R. J., Laiolo,
L., Lara-Lopez, A., Lozano-Montes, H., Matear, R. J., McEn-
nulty, F., Robson, B., Rochester, W., Skerratt, J., Smith, J. A.,
Strzelecki, J., Suthers, I. M., Swadling, K. M., van Ruth, P., and
Richardson, A. J.: Modeling what we sample and sampling what
we model: Challenges for zooplankton model assessment, Front.
Mar. Sci., 4, 1–19,,
Fasham, M. J. R., Ducklow, H. W., and McKelvie, S. M.: A
nitrogen-based model of plankton dynamics in the ocean mixed
layer, J. Mar. Res., 48, 591–639, 1990.
Fennel, K., Wilkin, J., Levin, J., Moisan, J., O’Reilly, J., and Haid-
vogel, D.: Nitrogen cycling in the Middle Atlantic Bight: Re-
sults from a three-dimensional model and implications for the
North Atlantic nitrogen budget, Global Biogeochem. Cy., 20, 1–
14,, 2006.
Fennel, K., Hetland, R., Feng, Y., and DiMarco, S.: A cou-
pled physical-biological model of the Northern Gulf of
Mexico shelf: model description, validation and analysis
of phytoplankton variability, Biogeosciences, 8, 1881–1899,, 2011.
Follows, M. J., Dutkiewicz, S., Grant, S., and Chisholm,
S. W.: Emergent biogeography of microbial commu-
nities in a model ocean., Science, 315, 1843–1846,, 2007.
Forristall, G. Z., Schaudt, K. J., and Cooper, C. K.: Evo-
lution and kinematics of a loop current eddy in the
Gulf of Mexico during 1985, J. Geophys. Res., 97, 2173,, 1992.
Franks, P. J. S.: NPZ models of plankton dynamics: Their construc-
tion, coupling to physics, and application, J. Oceanogr., 58, 379–
387,, 2002.
Franks, P. J. S.: Planktonic ecosystem models: Perplexing parame-
terizations and a failure to fail, J. Plankton Res., 31, 1299–1306,, 2009.
Geers, T. M., Pikitch, E. K., and Frisk, M. G.: An original model
of the northern Gulf of Mexico using Ecopath with Ecosim
and its implications for the effects of fishing on ecosystem
structure and maturity, Deep-Sea Res. Pt. II, 129, 319–331,, 2016.
Geider, R. J., Maclntyre, H. L., and Kana, T. M.: A dy-
namic regulatory model of phytoplanktonic acclimation to light,
nutrients, and temperature, Limnol. Oceangr., 43, 679–694,, 1998.
Gentleman, W., Leising, A., Frost, B., Strom, S., and Mur-
ray, J.: Functional responses for zooplankton feeding on
multiple resources: a review of assumptions and bio-
logical dynamics, Deep-Sea Res. Pt. II, 50, 2847–2875,, 2003a.
Gentleman, W., Leising, A., Frost, B., Strom, S., and Mur-
ray, J.: Functional responses for zooplankton feeding on
multiple resources: A review of assumptions and bio-
logical dynamics, Deep-Sea Res. Pt. II, 50, 2847–2875,, 2003b.
Gentleman, W. C. and Neuheimer, A. B.: Functional responses and
ecosystem dynamics: How clearance rates explain the influence
of satiation, food-limitation and acclimation, J. Plankton Res.,
30, 1215–1231,, 2008.
Gomez, F. A., Lee, S.-K., Liu, Y., Hernandez Jr., F. J., Muller-
Karger, F. E., and Lamkin, J. T.: Seasonal patterns in phyto-
plankton biomass across the northern and deep Gulf of Mex-
ico: a numerical model study, Biogeosciences, 15, 3561–3576,, 2018.
Gregg, W. W., Ginoux, P., Schopf, P. S., and Casey, N. W.: Phy-
toplankton and iron: Validation of a global three-dimensional
ocean biogeochemical model, Deep-Sea Res. Pt. II, 50, 3143–
3169,, 2003.
Gutiérrez-Rodríguez, A., Décima, M., Popp, B. N., and Landry,
M. R.: Isotopic invisibility of protozoan trophic steps in
marine food webs, Limnol. Oceanogr., 59, 1590–1598,, 2014.
Hill, H., Hill, C., Follows, M., and Dutkiewicz, S.: Is there a com-
putational advantage to offline tracer modelling at very high res-
olution?, Geophis. Res. Abstr., 6, 2005.
Holl, C. M., Waite, A. M., Pesant, S., Thompson, P. A., and Mon-
toya, J. P.: Unicellular diazotrophy as a source of nitrogen to
Leeuwin Current coastal eddies, Deep-Sea Res. Pt. II, 54, 1045–
1054,, 2007.
Hu, C., Lee, Z., and Franz, B.: Chlorophyll a algorithms for
oligotrophic oceans: A novel approach based on three-band
reflectance difference, J. Geophys. Res.-Oceans, 117, 1–25,, 2012.
Ikeda, T., Kanno, Y., Ozaki, K., and Shinada, A.: Metabolic
rates of epipelagic marine copepods as a function of Biogeosciences, 17, 3385–3407, 2020
3406 T. A. Shropshire et al.: Quantifying spatiotemporal variability in zooplankton dynamics
body mass and temperature, Mar. Biol., 139, 587–596,, 2001.
Kishi, M. J., Kashiwai, M., Ware, D. M., Megrey, B. a., Eslinger,
D. L., Werner, F. E., Noguchi-Aita, M., Azumaya, T., Fujii,
M., Hashimoto, S., Huang, D., Iizumi, H., Ishida, Y., Kang,
S., Kantakov, G. a., Kim, H. C., Komatsu, K., Navrotsky, V.
V., Smith, S. L., Tadokoro, K., Tsuda, A., Yamamura, O., Ya-
manaka, Y., Yokouchi, K., Yoshie, N., Zhang, J., Zuenko, Y. I.,
and Zvalinsky, V. I.: NEMURO-a lower trophic level model for
the North Pacific marine ecosystem, Ecol. Model., 202, 12–25,, 2007.
Kitchens, L. L. and Rooker, J. R.: Habitat associations of dolphin-
fish larvae in the Gulf of Mexico, Fish. Oceanogr., 23, 460–471,, 2014.
Kjellerup, S., Dünweber, M., Swalethorp, R., Nielsen, T. G., Møller,
E. F., Markager, S., and Hansen, B. W.: Effects of a future
warmer ocean on the coexisting copepods Calanus finmarchicus
and C. glacialis in Disko Bay, western Greenland, Mar. Ecol.-
Prog. Ser., 447, 87–108,,
Landry, M., Haas, L., and Fagerness, V.: Dynamics of
microbial plankton communities: experiments in Ka-
neohe Bay, Hawaii, Mar. Ecol.-Prog. Ser., 16, 127–133,, 1984.
Landry, M. R. and Calbet, A.: Microzooplankton produc-
tion in the oceans, ICES J. Mar. Sci., 61, 501–507,, 2004.
Landry, M. R. and Hassett, R. P.: Estimating the grazing im-
pact of marine micro-zooplankton, Mar. Biol., 67, 283–288,, 1982.
Landry, M. R., Decima, M., Simmons, M. P., Hannides,
C. C. S., and Daniels, E.: Mesozooplankton biomass
and grazing responses to Cyclone Opal, a subtropical
mesoscale eddy, Deep-Sea Res. Pt. II, 55, 1378–1388,, 2008.
Landry, M. R., Ohman, M. D., Goericke, R., Stukel, M. R.,
and Tsyrklevich, K.: Lagrangian studies of phytoplankton
growth and grazing relationships in a coastal upwelling ecosys-
tem off Southern California, Prog. Oceanogr., 83, 208–216,, 2009.
Landry, M. R., Selph, K. E., Decima, M., Gutierrez-
Rodríguez, A., Stukel, M. R., Taylor, A. G., and Pasulka,
A. L.: Phytoplankton production and grazing balances
in the Costa Rica Dome, J. Plankton Res., 38, 366–379,, 2016.
Landry, M. R., Beckley, L. E., and Muhling, B. A.: Climate sensi-
tivities and uncertainties in food-web pathways supporting larval
bluefin tuna in subtropical oligotrophic oceans, ICES J. Mar. Sci.,
76, 359–369,, 2019.
Large, W. G., McWilliams, J. C., and Doney, S. C.: Oceanic
vertical mixing: A review and a model with a nonlocal
boundary layer parameterization, Rev. Geophys., 32, 363–403,, 1994.
Laxenaire, R., Speich, S., Blanke, B., Chaigneau, A.,
Pegliasco, C., and Stegner, A.: Anticyclonic Eddies
Connecting the Western Boundaries of Indian and At-
lantic Oceans, J. Geophys. Res.-Oceans, 123, 7651–7677,, 2018.
Laxenaire, R., Speich, S., and Stegner, A.: Evolution of the Thermo-
haline Structure of One Agulhas Ring Reconstructed from Satel-
lite Altimetry and Argo Floats, J. Geophys. Res.-Oceans, 124,
8969–9003,, 2019.
Li, Q. P., Franks, P. J. S., Landry, M. R., Goericke, R.,
and Taylor, A. G.: Modeling phytoplankton growth rates
and chlorophyll to carbon ratios in California coastal and
pelagic ecosystems, J. Geophys. Res.-Biogeo., 115, 1–12,, 2010.
Lindo-Atichati, D., Bringas, F., Goni, G., Muhling, B.,
Muller-Karger, F. E., and Habtes, S.: Varying mesoscale
structures influence larval fish distribution in the north-
ern Gulf of Mexico, Mar. Ecol.-Prog. Ser., 463, 245–257,, 2012.
The Relationship between Variations in the Gulf of
Mexico Loop Current and Straits of Florida Volume
Transport 23, 785–796,
0485(1993)023<0785:TRBVIT>2.0.CO;2, 1993.
McKinley, G. A., Follows, M. J., and Marshall, J.: Mecha-
nisms of air-sea CO2flux variability in the equatorial Pacific
and the North Atlantic, Global Biogeochem. Cy., 18, 1–14,, 2004.
Mitra, A., Flynn, K. J., Burkholder, J. M., Berge, T., Calbet,
A., Raven, J. A., Granéli, E., Glibert, P. M., Hansen, P. J.,
Stoecker, D. K., Thingstad, F., Tillmann, U., Våge, S., Wilken,
S., and Zubkov, M. V.: The role of mixotrophic protists in
the biological carbon pump, Biogeosciences, 11, 995–1005,, 2014.
Moeller, H. V., Laufkötter, C., Sweeney, E. M., and Johnson,
M. D.: Light-dependent grazing can drive formation and deep-
ening of deep chlorophyll maxima, Nat. Commun., 10, 1–8,, 2019.
Morey, S. L., Martin, P. J., O’Brien, J. J., Wallcraft, A. A., and
Zavala-Hidalgo, J.: Export pathways for river discharged fresh
water in the northern Gulf of Mexico, J. Geophys. Res.-Oceans,
108, 1–1,, 2003a.
Morey, S. L., Schroeder, W. W., O’Brien, J. J., and Zavala-
Hidalgo, J.: The annual cycle of riverine influence in the east-
ern Gulf of Mexico basin, Geophys. Res. Lett., 30, 1867,, 2003b.
Morey, S. L., Zavala-Hidalgo, J., and O’Brien, J. J.: The Sea-
sonal Variability of Continental Shelf Circulation in the Northern
and Western Gulf of Mexico from a High-Resolution Numerical
Model, in Circulation in the Gulf of Mexico: Observations and
Models, GMS, 161, 203–218, 2013.
Moriarty, R. and O’Brien, T. D.: Distribution of mesozooplankton
biomass in the global ocean, Earth Syst. Sci. Data, 5, 45–55,, 2013.
Morrow, R. M., Ohman, M. D., Goericke, R., Kelly, T. B., Stephens,
B. M., and Stukel, M. R.: CCE V: Primary production, mesozoo-
plankton grazing, and the biological pump in the California Cur-
rent Ecosystem: Variability and response to El Niño, Deep-Sea
Res. Pt. I, 140, 52–62,,
Muhling, B. A., Lamkin, J. T., Alemany, F., García, A., Farley, J.,
Ingram, G. W., Berastegui, D. A., Reglero, P., and Carrion, R.
L.: Reproduction and larval biology in tunas, and the importance
of restricted area spawning grounds, Rev. Fish Biol. Fish., 27,
697–732,, 2017.
Biogeosciences, 17, 3385–3407, 2020
T. A. Shropshire et al.: Quantifying spatiotemporal variability in zooplankton dynamics 3407
Mulholland, M. R., Bernhardt, P. W., Heil, C. A., Bronk, D. A., and
O’Neil, J. M.: Nitrogen fixation and release of fixed nitrogen by
Trichodesmium spp. in the Gulf of Mexico, Limnol. Oceanogr.,
51, 2484,, 2006.
Murray, A. G. and Parslow, J. S.: The analysis of alternative formu-
lations in a simple model of a coastal ecosystem, Ecol. Model.,
119, 149–166,,
Oey, L., Ezer, T., and Lee, H.: Loop Current, rings and related cir-
culation in the Gulf of Mexico: A review of numerical models
and future challenges, Geophys. Monogr., 161, 31–56, 2005.
Parker, R. A.: Dynamic models for ammonium inhibition of ni-
trate uptake by phytoplankton, Ecol. Model., 66, 113–120,, 1993.
Pegliasco, C., Chaigneau, A., and Morrow, R.: Main eddy ver-
tical structures observed in the four major Eastern Boundary
Upwelling Systems, J. Geophys. Res.-Oceans, 120, 6008–6033,, 2015.
Platt, T., Gallegos, C. L., and Harrison, W. G.: Photoinhibition of
photosynthesis in natural assemblages of marine phytoplankton,
J. Mar. Res., 38, 687–701, 1980.
Pörtner, H. O. and Farrell, A. P.: Physiology and Climate Change
Hans, Science, 322, 690–692, 2008.
Richardson, A. J.: In hot water: Zooplankton and cli-
mate change, ICES J. Mar. Sci., 65, 279–295,, 2008.
Riley, G. A.: Factors controlling phytoplankton populations on
Georges Bank, J. Mar. Res., 6, 54–73, 1946.
Rooker, J. R., Simms, J. R., David Wells, R. J., Holt, S. A.,
Holt, G. J., Graves, J. E., and Furey, N. B.: Distribution
and habitat associations of billfish and swordfish larvae across
mesoscale features in the gulf of Mexico, PLoS One, 7, e34180,, 2012.
Rooker, J. R., Kitchens, L. L., Dance, M. A., Wells, R. J. D., Falter-
man, B., and Cornic, M.: Spatial, Temporal, and Habitat-Related
Variation in Abundance of Pelagic Fishes in the Gulf of Mexico:
Potential Implications of the Deepwater Horizon Oil Spill, PLoS
One, 8, e76080,,
Sailley, S. F., Vogt, M., Doney, S. C., Aita, M. N., Bopp, L., Buiten-
huis, E. T., Hashioka, T., Lima, I., Le Quéré, C., and Yamanaka,
Y.: Comparing food web structures and dynamics across a suite
of global marine ecosystem models, Ecol. Model., 261–262, 43–
57,, 2013.
Sailley, S. F., Polimene, L., Mitra, A., Atkinson, A., and Allen,
J. I.: Impact of zooplankton food selectivity on plankton dy-
namics and nutrient cycling, J. Plankton Res., 37, 519–529,, 2015.
Selph, K. E., Landry, M. R., Taylor, A. G., Gutierrez-
Rodríguez, A., Stukel, M. R., Wokuluk, J., and Pasulka,
A.: Phytoplankton production and taxon-specific growth rates
in the Costa Rica Dome, J. Plankton Res., 38, 199–215,, 2016.
Sherr, E. B. and Sherr, B. F.: Significance of predation
by protists in aquatic microbial food webs, Antonie van
Leeuwenhoek, Int. J. Gen. Mol. Microbiol., 81, 293–308,, 2002.
Shropshire, T.: NEMURO-GoM model code along with configu-
ration files to run with MITgcm, GitHub, available at: https:
//, 2019a.
Shropshire, T.: Temporally averaged three-dimensional fields for all
11 state variables used in NEMURO-GOM along with surface
chlorophyll, available at:,
Staniewski, M. A. and Short, S. M.: Methodological review and
meta-analysis of dilution assays for estimates of virus- and
grazer-mediated phytoplankton mortality, Limnol. Oceanogr.-
Methods, 16, 649–668,,
Steele, J. H. and Henderson, E. W.: The role of preda-
tion in plankton models, J. Plankton Res., 14, 157–172,, 1992.
Steinberg, D. K. and Landry, M. R.: Zooplankton and the
Ocean Carbon Cycle, Annu. Rev. Mar. Sci., 9, 413–444,, 2017.
Straile, D.: and metazoan efficiencies of protozoan Gross growth
on food concentration, and their dependence zooplankton group
ratio, and taxonomic, Limnol. Oceanogr., 42, 1375–1385, 1997.
Strickland, J. D. H. and Parsons, T. R.: A practical handbook
for seawater analysis. Second Edition, available at: http://www. (last access: 29 June 2020),
Strom, S. L., Benner, R., Ziegler, S., and Dagg, M. J.: Plank-
tonic grazers are a potentially important source of marine
dissolved organic carbon, Limnol. Oceanogr., 42, 1364–1374,, 1997.
Stukel, M. R., Coles, V. J., Brooks, M. T., and Hood, R. R.: Top-
down, bottom-up and physical controls on diatom-diazotroph as-
semblage growth in the Amazon River plume, Biogeosciences,
11, 3259–3278,, 2014.
Stukel, M. R., Kahru, M., Benitez-Nelson, C. R., Décima,
M., Goericke, R., Landry, M. R., and Ohman, M. D.: Us-
ing Lagrangian-based process studies to test satellite algo-
rithms of vertical carbon flux in the eastern North Pa-
cific Ocean, J. Geophys. Res.-Oceans, 120, 7208–7222,, 2015.
Turner, J. T.: Zooplankton fecal pellets, marine snow, phytodetritus
and the ocean’s biological pump, Prog. Oceanogr., 130, 205–248,, 2015.
Wainwright, T. C., Feinberg, L. R., Hooff, R. C., and Peter-
son, W. T.: A comparison of two lower trophic models for
the California Current System, Ecol. Modell., 202, 120–131,, 2007.
Werner, F. E., Ito, S. I., Megrey, B. A., and Kishi, M. J.: Syn-
thesis of the NEMURO model studies and future directions
of marine ecosystem modeling, Ecol. Model., 202, 211–223,, 2007.
Wiebe, P. H.: Functional regression equations for zooplankton dis-
placement volume wet weight, dry weight, and carbon: A correc-
tion, Fish. Bull., 86, 833–835, 1988.
Xue, Z., He, R., Fennel, K., Cai, W.-J., Lohrenz, S., and Hopkin-
son, C.: Modeling ocean circulation and biogeochemical vari-
ability in the Gulf of Mexico, Biogeosciences, 10, 7219–7234,, 2013. Biogeosciences, 17, 3385–3407, 2020
... Recent advances in formal assimilation of biogeochemical properties into ocean models are beginning to allow objective model parameterization, a crucial first step for treating models as testable hypotheses (Xiao and Friedrichs, 2014a;Mattern and Edwards, 2019;Kaufman et al., 2018;Ford et al., 2018;Kriest et al., 2017;Shen et al., 2016;Oschlies, 2006;DeVries and Weber, 2017;Nowicki et al., 2022). Nevertheless, most of these approaches rely only on the assimilation of surface chlorophyll and/or other phytoplankton properties, thus leading to potentially high inaccuracies in parameterizing zooplankton dynamics (Shropshire et al., 2020;Löptien and Dietze, 2015). This is particularly important, because inaccurate parameterizations of mesozooplankton may lead to qualitatively and quantitatively inaccurate export dynamics (Cavan et al., 2017;Anderson et al., 2013). ...
... 2.1 Core NEMURO BCP model NEMURO BCP was developed from the NEMURO class of models originally developed for the North Pacific (Kishi et al., , 2011Yoshie et al., 2007) and includes several modifications adapted by Shropshire et al. (2020) that allow the model to be compared more directly to common field measurements. It also includes three optional modules that can be toggled on and off (the carbon system, nitrogen isotopes, and 234 Th). ...
... The core of NEMURO BCP is nitrogen-based and includes 19 state variables (Table 1): three nutrients (nitrate, ammonium, and silicic acid), two phytoplankton (small phytoplankton and diatoms), five zooplankton (protistan zooplankton, small non-vertically migrating mesozooplankton, small vertically migrating mesozooplankton, large nonvertically migrating mesozooplankton, large vertically migrating mesozooplankton), two dissolved organic pools (labile dissolved organic nitrogen and refractory dissolved organic nitrogen), four non-living particulate pools (small par-ticulate nitrogen, large particulate nitrogen, small opal, and large opal), two chlorophyll state variables (one associated with small phytoplankton, the other with diatoms), and oxygen. As in Shropshire et al. (2020), the small and large mesozooplankton are defined based on size (< 1 and > 1 mm, respectively) rather than trophic status to allow direct comparison to common size-fractionated measurements. Relative to the original NEMURO model, key changes include (1) an explicit chlorophyll module (based on Li et al., 2010) that allows direct comparison to in situ chlorophyll measurements and gut pigment measurements made with herbivorous zooplankton; (2) division of dissolved organic matter into refractory and labile dissolved organic nitrogen to simulate subduction of refractory molecules; (3) division of detrital pools into slowly and rapidly sinking particles to simulate more accurately the gravitational pump; (4) division of mesozooplankton into epipelagic-resident taxa and vertical migrants to simulate active transport by diel vertical migrators; and (5) addition of a dissolved oxygen state variable that potentially limits heterotrophic growth in the mesopelagic ocean. ...
Full-text available
The ability to constrain the mechanisms that transport organic carbon into the deep ocean is complicated by the multiple physical, chemical, and ecological processes that intersect to create, transform, and transport particles in the ocean. In this paper we develop and parameterize a data-assimilative model of the multiple pathways of the biological carbon pump (NEMUROBCP). The mechanistic model is designed to represent sinking particle flux, active transport by vertically migrating zooplankton, and passive transport by subduction and vertical mixing, while also explicitly representing multiple biological and chemical properties measured directly in the field (including nutrients, phytoplankton and zooplankton taxa, carbon dioxide and oxygen, nitrogen isotopes, and 234Thorium). Using 30 different data types (including standing stock and rate measurements related to nutrients, phytoplankton, zooplankton, and non-living organic matter) from Lagrangian experiments conducted on 11 cruises from four ocean regions, we conduct an objective statistical parameterization of the model and generate 1 million different potential parameter sets that are used for ensemble model simulations. The model simulates in situ parameters that were assimilated (net primary production and gravitational particle flux) and parameters that were withheld (234Thorium and nitrogen isotopes) with reasonable accuracy. Model results show that gravitational flux of sinking particles and vertical mixing of organic matter from the euphotic zone are more important biological pump pathways than active transport by vertically migrating zooplankton. However, these processes are regionally variable, with sinking particles most important in oligotrophic areas of the Gulf of Mexico and California Current, sinking particles and vertical mixing roughly equivalent in productive coastal upwelling regions and the subtropical front in the Southern Ocean, and active transport an important contributor in the eastern tropical Pacific. We further find that mortality at depth is an important component of active transport when mesozooplankton biomass is high, but it is negligible in regions with low mesozooplankton biomass. Our results also highlight the high degree of uncertainty, particularly amongst mesozooplankton functional groups, that is derived from uncertainty in model parameters. Indeed, variability in BCP pathways between simulations for a specific location using different parameter sets (all with approximately equal misfit relative to observations) is comparable to variability in BCP pathways between regions. We discuss the implications of these results for other data-assimilation approaches and for studies that rely on non-ensemble model outputs.
... Surface velocity fields were provided by the reanalysis data product OSCAR (ESR, 2009), and the floats were tracked backward in time using a 4th-order Runge-Kutta algorithm with a 1-h time step and a model duration of ∼8 weeks total (April-June), or 6 weeks for each individual experiment. Subgrid scale dispersion was parameterized with a random-walk perturbation corresponding to a horizontal eddy diffusivity of 20 m 2 s −1 , a value consistent with previous regional biogeochemical models (Shropshire et al., 2020). This approach assumes that the floats stay within the surface mixed layer, which may not be representative of actual particle behaviors in convergence or divergence zones. ...
... Knapp et al. (2022) apply a δ 15 N budgeting approach to evaluate the relative importance of subsurface NO 3 − and N 2 fixation for supporting net food web process and export in the oceanic GoM. In additional synthetic studies presented elsewhere, Shropshire et al. (2020) configure a physical-biogeochemical model (NEMURO-GoM) that is validated against mesozooplankton biomass and grazing rates and other constraints from the BLOOFINZ cruises, and Kelly et al. (2021) evaluate the contribution of laterally sourced organic matter to export using the NEMURO-GoM model and satellite remote sensing products. Significant findings from the dynamics and flux portion of the study include the following: ...
... Grazing on phytoplankton was low (1-3% of Chla consumed d −1 ), but estimated trophic fluxes from microzooplankton consumption and carnivory were sufficient to satisfy the C demand for respiration and growth of suspension-feeding mesozooplankton estimated from empirical relationships. • A NEMURO-based (Kishi et al., 2007) physicalbiogeochemical model parameterized with BLOOFINZ-GoM rate and relationship data captured broad ecosystem attributes including phytoplankton and mesozooplankton biomass, depth of the DCM and nutricline, and growth and grazing patterns (Shropshire et al., 2020). Regional mesozooplankton production estimated from the model average 66 ± 8 × 10 9 kg C y −1 . ...
Western Atlantic bluefin tuna (ABT) undertake long-distance migrations from rich feeding grounds in the North Atlantic to spawn in oligotrophic waters of the Gulf of Mexico (GoM). Stock recruitment is strongly affected by interannual variability in the physical features associated with ABT larvae, but the nutrient sources and food-web structure of preferred habitat, the edges of anticyclonic loop eddies, are unknown. Here, we describe the goals, physical context, design and major findings of an end-to-end process study conducted during peak ABT spawning in May 2017 and 2018. Mesoscale features in the oceanic GoM were surveyed for larvae, and five multi-day Lagrangian experiments measured hydrography and nutrients; plankton biomass and composition from bacteria to zooplankton and fish larvae; phytoplankton nutrient uptake, productivity and taxon-specific growth rates; micro- and mesozooplankton grazing; particle export; and ABT larval feeding and growth rates. We provide a general introduction to the BLOOFINZ-GoM project (Bluefin tuna Larvae in Oligotrophic Ocean Foodwebs, Investigation of Nitrogen to Zooplankton) and highlight the finding, based on backtracking of experimental waters to their positions weeks earlier, that lateral transport from the continental slope region may be more of a key determinant of available habitat utilized by larvae than eddy edges per se.
... mm range), observed mesozooplankton biomass in the 0.2-1.0 mm range (herein referred to large zooplankton biomass, LZB) was adjusted to the top 25 m of the water column and scaled using the ratio of SZB to LZB estimated by a threedimensional biogeochemical model NEMURO-GoM (Shropshire et al., 2020). The NEMURO-GoM model was designed to simulate zooplankton biomass distribution in the GoM and has been extensively validated using remote and in situ measurements including over two decades of zooplankton biomass measurements collected by the Southeast Area Monitoring and Assessment Program (SEAMAP). ...
... An advantage of the LZB and SZB was that they could be estimated for each ABT larval sampling location, as bulk zooplankton net tows did not accompany every larval net tow. For more information on NEMURO-GoM see Shropshire et al. (2020). Two additional variables were calculated from SZB and LZB that estimated the respective prey biomasses defined as a function of ABT-larval length for small prey biomass (SPB) and for large prey biomass (LPB), respectively (see Table IV). ...
Larval abundances of Atlantic bluefin tuna (ABT) in the Gulf of Mexico are currently utilized to inform future recruitment by providing a proxy for the spawning potential of western ABT stock. Inclusion of interannual variations in larval growth is a key advance needed to translate larval abundance to recruitment success. However, little is known about the drivers of growth variations during the first weeks of life. We sampled patches of western ABT larvae in 3–4 day Lagrangian experiments in May 2017 and 2018, and assessed age and growth rates from sagittal otoliths relative to size categories of zooplankton biomass and larval feeding behaviors from stomach contents. Growth rates were similar, on average, between patches (0.37 versus 0.39 mm d−1) but differed significantly through ontogeny and were correlated with a food limitation index, highlighting the importance of prey availability. Otolith increment widths were larger for postflexion stages in 2018, coincident with high feeding on preferred prey (mainly cladocerans) and presumably higher biomass of more favorable prey type. Faster growth reflected in the otolith microstructures may improve survival during the highly vulnerable larval stages of ABT, with direct implications for recruitment processes.
... In estuarine environments, zooplankton species diversity and abundance can be affected by many physicochemical conditions and food availability (Froneman, 2004;Modeŕan et al., 2010;Nandy and Mandal, 2020;Shropshire et al., 2020;Telesh, 2004). In particular, the impact of suspended particulate matter (SPM) on plankton food availability in estuarine-coastal environments may be slightly controversial. ...
Full-text available
In temperate estuaries, rainfall causes environmental fluctuations, such as salinity and suspended particulate matter (SPM), and can affect zooplankton distribution patterns. This study focused on the effect of temporary freshwater inflow on the composition of major zooplankton species and changes in their ecological status in the Seomjin River estuary before (June) and after (August) rainfall in 2018. Environmental data were collected from 14 and 15 stations before and after rainfall, respectively. All factors except for chlorophyll-a (Chla) concentration differed significantly before and after rainfall p<0.05), and a salinity gradient extended to Yeosu Bay from Gwangyang Bay. Zooplankton abundance decreased significantly after rainfall. There was a high correlation between indicator species abundance and environmental factors after rainfall (correlation coefficient: 0.7521); however, the indicator species and environmental factors did not exhibit a significant correlation with salinity before rainfall. In terms of feeding habit composition, the carnivore proportion showed a significant decrease after rainfall compared to before rainfall (p<0.001), while the particle feeder proportion showed a significant increase after rainfall compared to before (p<0.001). In particular, Corycaeus spp. contributed significantly to the decrease in carnivore abundance after rainfall. Among the particle feeders, Copepodites significantly increased in abundance after rainfall. Carnivore abundance was negatively correlated with salinity, and particle feeder abundance was positively correlated with potential prey sources (SPM and Chl-a concentration), suggesting that particle feeders respond to the food-rich environment after rainfall.
... Given that isopycnal shallowing or deepening due to either CEs or AEs increases or decreases the NN stock, relationships between physical variables and the NN stock can be used to calculate the vertical distribution of NN. In the GM, both univariate linear parameterizations and complex biogeochemical models have been used to determine the NN concentration (Jolliff et al., 2008;Pasqueron de Fommervault et al., 2017;Damien et al., 2018;Gomez et al., 2018;Shropshire et al., 2020). The biogeochemical models are able to reproduce the characteristics of vertical nitrate profiles (Jolliff et al., 2008;Damien et al., 2018;Gomez et al., 2018), although they commonly underestimate nitrate concentrations in waters from 300-500 m. ...
Full-text available
In the ocean, nitrogen availability is an important control of primary production and influences the amount of energy flowing through food webs. Mesoscale eddies play important roles in modulating the spatial distributions of physical and biogeochemical properties in the Gulf of Mexico (GM), including the availability of nitrate + nitrite (NN). In this study, we explore an oceanographic station classification based on the integrated NN stock that we have named the “nitracentric classification” and a classification based on hydrographic variables that we call the Best Fit Variables (BFVs), such as the depth of the 20°C isotherm and the depth of the 26 kg m-3 isopycnal, to identify stations under the influence of mesoscale eddies. We analyzed hydrographic profiles of CTD data and the NN concentrations in discrete samples collected in June 2016 during the oceanographic campaign XIXIMI-5, which was conducted in the deep-water region of the GM. The best station separation was produced when the NN concentration was integrated between the surface and 200 m depth, which was supported by the station classification based on the BFVs. Our classification system produces a better separation between station groups when compared to other classifications that rely on the use of altimetric variables and hydrographic criteria that have been previously employed to study biogeochemical and physical processes in the GM. We obtained parameterizations that accurately predicted the NN profiles between 100–500 m of stations sampled under stratified conditions in two other XIXIMI cruises in the gulf, although the parameterization has to be adapted to obtain accurate predictions under winter mixing conditions. Our results can be used to predict nitrate stocks and profiles based on a single BFV value obtained from the existing hydrographic databases of the GM as well as from CTD data at the time of sampling. The analysis of the CLIVAR Section A22 in the Caribbean Sea indicates that the nitracentric and hydrographic classification methodology developed in this study can also be applied to other oligotrophic basins where mesoscale eddies play important roles in controlling the distributions of hydrographic and biogeochemical properties.
... For the Gulf of Mexico, peak occurrence and observations were clustered off the Mississippi River delta, an area of known high concentrations of large zooplankton (see Figs. 3 and 7 in 61 ). Nearshore habitats in the Gulf of Mexico may contain important food sources for manta rays 61 . Similarly, zooplankton biomass estimates 62 suggest high concentrations of potential manta ray prey in the Mississippi River plume, Florida coastal waters, the upwelling zone near Cape Hatteras, North Carolina, and the northeastern United States shelf-edge areas covered by the NYSERDA surveys. ...
Full-text available
In 2018, the giant manta ray was listed as threatened under the U.S. Endangered Species Act. We integrated decades of sightings and survey effort data from multiple sources in a comprehensive species distribution modeling (SDM) framework to evaluate the distribution of giant manta rays off the eastern United States, including the Gulf of Mexico. Manta rays were most commonly detected at productive nearshore and shelf-edge upwelling zones at surface thermal frontal boundaries within a temperature range of approximately 20–30 °C. SDMs predicted highest nearshore occurrence off northeastern Florida during April, with the distribution extending northward along the shelf-edge as temperatures warm, leading to higher occurrences north of Cape Hatteras, North Carolina from June to October, and then south of Savannah, Georgia from November to March as temperatures cool. In the Gulf of Mexico, the highest nearshore occurrence was predicted around the Mississippi River delta from April to June and again from October to November. SDM predictions will allow resource managers to more effectively protect manta rays from fisheries bycatch, boat strikes, oil and gas activities, contaminants and pollutants, and other threats.
... Recently, Jahn et al. (2019) developed a coupled physical-ecosystem model, which has quantified skill in capturing the intraseasonal Chl concentration signal. Several studies have reported the fidelity of the distribution of Chl concentrations simulated by the model (e.g., Kuhn et al., 2019;McParland & Levine, 2019;Shropshire et al., 2020;Tréguer et al., 2018). The motivation for this paper is to examine the mechanism of the CIO mode impact on the Chl concentration variability in the mixed layer during boreal summer (June-September; JJAS). ...
Full-text available
This study diagnoses the drivers of intraseasonal variability of mixed layer chlorophyll‐a (Chl) concentration in the tropical Indian Ocean during boreal summer (June–September) using a well‐tested coupled ocean‐ecosystem model. Results show that wind forcing is the primary source for energetic intraseasonal Chl concentration variability in the eastern Arabian Sea and western Bay of Bengal (BoB) modulated by the Central Indian Ocean (CIO) mode. The atmospheric anomalous anticyclone associated with the positive phase of CIO mode drives southeastward currents in the Arabian Sea, which transport nutrient and Chl towards the east. Simultaneously, the wind‐induced offshore currents strengthen upwelling in the western boundary of BoB, resulting in enhanced Chl in the upper ocean. Conversely, because of the formation of a barrier layer, the nutrient supply is suppressed along the equator, which causes a negative correlation between the CIO mode and Chl. This study complements the Madden‐Julian Oscillation forcing on the intraseasonal Chl concentration anomaly in the tropical Indian Ocean.
... spanned several orders of magnitude for a particular parameter. We then defined an initial guess for each parameter based primarily on values used in other NEMURO modelsShropshire et al. 2020;Yoshie et al. 2007). We first ran 30-day simulations for all 49 Lagrangian experiments using the initial parameter values and ...
Full-text available
The ability to constrain the mechanisms that transport organic carbon into the deep ocean is complicated by the multiple physical, chemical, and ecological processes that intersect to create, transform, and transport particles in the ocean. In this manuscript we develop and parameterize a data-assimilative model of the multiple pathways of the biological carbon pump (NEMUROBCP). The mechanistic model is designed to represent sinking particle flux, active transport by vertically migrating zooplankton, and passive transport by subduction and vertical mixing, while also explicitly representing multiple biological and chemical properties measured directly in the field (including nutrients, phytoplankton and zooplankton taxa, carbon dioxide and oxygen, nitrogen isotopes, and 234Thorium). Using 30 different data types (including standing stock and rate measurements related to nutrients, phytoplankton, zooplankton, and non-living organic matter) from Lagrangian experiments conducted on 11 cruises from four ocean regions, we conduct an objective statistical parameterization of the model and generate one million different potential parameter sets that are used for ensemble model simulations. The model simulates in situ parameters that were assimilated (net primary production and gravitational particle flux) and parameters that were withheld (234Thorium and nitrogen isotopes) with reasonable accuracy. Model results show that gravitational flux of sinking particles and vertical mixing of organic matter from the surface ocean are more important biological pump pathways than active transport by vertically-migrating zooplankton. However, these processes are regionally variable, with sinking particles most important in oligotrophic areas of the Gulf of Mexico and California, sinking particles and vertical mixing roughly equivalent in productive regions of the CCE and the subtropical front in the Southern Ocean, and active transport an important contributor in the Eastern Tropical Pacific. We further find that mortality at depth is an important component of active transport when mesozooplankton biomasses are high, but that it is negligible in regions with low mesozooplankton biomass. Our results also highlight the high degree of uncertainty, particularly amongst mesozooplankton functional groups, that is derived from uncertainty in model parameters, with important implications from results that rely on non-ensemble model outputs. We also discuss the implications of our results for other data assimilation approaches.
... Moreover, due to their small size, limited capacity for self-dispersal, and high sensitivity to environmental change, zooplankton are also useful bioindicators when evaluating the health of marine ecosystems (Buttay et al., 2015;Johnston et al., 2015;Parmar et al., 2016;Yang & Zhang, 2019). Environmental change and ecological processes shape zooplankton communities, both in terms of their spatial and temporal abundance (Färber Lorda et al., 2019;Shropshire et al., 2020) and taxonomic composition (Buttay et al., 2015;Cruz-Rosado et al., 2020), and may lead to the loss of zooplankton biodiversity (Gazonato Neto et al., 2014;Parmar et al., 2016), which can affect ecosystem services and result in economic consequences (Beaugrand et al., 2010;Bucklin et al., 2016;Everaert et al., 2018;Johnston et al., 2015). ...
Zooplankton plays a pivotal role in sustaining the majority of marine ecosystems. The distribution patterns and diversity of zooplankton provide key information for understanding the functioning of these ecosystems. Nevertheless, due to the numerous cryptic and sibling species and the lack of diagnostic characteristics for early developmental stages, the identification of the global-to-local patterns of zooplankton biodiversity and biogeography remains challenging in different research fields. The spatial and temporal changes in the zooplankton community in the open waters of the southern Gulf of Mexico were assessed using metabarcoding analysis of the V9 region of 18S rRNA and mitochondrial cytochrome oxidase c subunit I (COI). Additionally, a multi-scale analysis was implemented to evaluate which environmental predictors may explain the variability in the structure of the zooplankton community. Our findings suggest that the synergistic effects of dissolved oxygen concentration, temperature, and longitude (intended as a proxy for still unidentified predictors) may explain both spatial and temporal zooplankton variability even with low contribution. Furthermore, the zooplankton distribution likely reflects the coexistence of three heterogeneous ecoregions and a bio-physical partitioning of the studied area. Finally, some taxa were either exclusive or predominant with either 18S or COI markers. This may suggest that comprehensive assessments of the zooplankton community may be more accurately met by the use of multi-locus approaches.
Nutrient limitation on phytoplankton growth plays a critical role in ocean productivity, the functioning of marine ecosystems, and the ocean carbon cycle. In the Celtic Sea, a temperate shelf sea, many studies have shown the importance of nitrate on phytoplankton growth focusing on the seasonal cycle of nitrate and feedbacks with the physical environment; but only recently has it been demonstrated, through discrete measurements, that dissolved iron also plays an important role in the ecosystem of the region. A well-established one-dimensional model has been developed to analyse the nutrient co-limitation between dissolved iron and dissolved inorganic nitrogen in the Celtic Sea. This model allows us to study the full seasonal cycle and inter-annual variability of these two nutrients. Simulations show that dissolved iron is an important nutrient for the development of the spring bloom, while nitrate plays a more important role during the summer season. Sensitivity analyses show that these results are robust when varying the nutrient-related parameters; the largest variability observed for primary production was observed when varying the nutrient sediment flux rates for dissolved iron and nitrate while less impact on phytoplankton production occurs when changing the half saturation constants. Here, we demonstrate that dissolved iron is an important nutrient for the development of the spring bloom and it should not be neglected as a state variable when modelling the Celtic Sea or other temperate shelf seas.
Full-text available
The transfer of Indian Ocean thermocline and intermediate waters into the South Atlantic via the Agulhas leakage is generally believed to be primarily accomplished through mesoscale eddy processes, essentially anticyclones known as Agulhas Rings. Here we take advantage of a recent eddy tracking algorithm and Argo float profiles to study the evolution and the thermohaline structure of one of these eddies over the course of 1.5 years (May 2013–November 2014). We found that during this period the ring evolved according to two different phases: During the first one, taking place in winter, the mixing layer in the eddy deepened significantly. During the second phase, the eddy subsided below the upper warmer layer of the South Atlantic subtropical gyre while propagating west. The separation of this eddy from the sea surface could explain the decrease in its surface signature in satellite altimetry maps, suggesting that such changes are not due to eddy dissipation processes. It is a very large eddy (7.1×10^{13} m^{3} in volume), extending, after subduction, from a depth of 200–1,200 m and characterized by two mode water cores. The two mode water cores represent the largest eddy heat and salt anomalies when compared with the surrounding. In terms of its impact over 1 year, the north-westward propagation of this long-lived anticyclone induces a transport of 2.2 Sv of water, 0.008 PW of heat, and 2.2×10^{5} kg s^{−1} of salt. These results confirm that Agulhas Rings play a very important role in the Indo-Atlantic interocean exchange of heat and salt.
Full-text available
Deep Chlorophyll Maxima (DCMs) are subsurface peaks in chlorophyll-a concentration that may coincide with peaks in phytoplankton abundance and primary productivity. Work on the mechanisms underlying DCM formation has historically focused on phytoplankton physiology (e.g., photoacclimation) and behavior (e.g., taxis). While these mechanisms can drive DCM formation, they do not account for top-down controls such as predation by grazers. Here, we propose a new mechanism for DCM formation: Light-dependent grazing by microzooplankton reduces phytoplankton biomass near the surface but allows accumulation at depth. Using mathematical models informed by grazing studies, we demonstrate that light-dependent grazing is sufficient to drive DCM formation. Further, when acting in concert with other mechanisms, light-dependent grazing deepens the DCM, improving the fit of a global model with observational data. Our findings thus reveal another mechanism by which microzooplankton may regulate primary production, and impact our understanding of biogeochemical cycling at and above the DCM.
Full-text available
The Indo-Atlantic interocean exchanges achieved by Agulhas Rings are tightly linked to global ocean circulation and climate. Yet they are still poorly understood because they are difficult to identify and follow. We propose here an original assessment on Agulhas Rings, achieved by TOEddies, a new eddy identification and tracking algorithm that we applied over 24 years of satellite altimetry. Its main novelty lies in the detection of eddy splitting and merging events. These are particularly abundant and significantly impact the concept of a trajectory associated with a single eddy, which becomes less obvious than previously admitted. To overcome this complication, we have defined a network of segments that group together in relatively complex trajectories. Such a network provides an original assessment of the routes and the history of Agulhas Rings. It links 730,481 eddies into 6,363 segments that cluster into Agulhas Ring trajectories of different orders. Such an order depends on the affiliation of the eddies and segments, in a similar way as a tree of life. Among them, we have identified 122 order 0 trajectories that can be considered as the major trajectories associated to a single eddy, albeit it has undergone itself splitting and merging events. Despite the disappearance of many eddies in the altimeter signal in the Cape Basin, a significant fraction can be followed from the Indian Ocean to the South Brazil Current with, on average, 3.5 years to cross the entire South Atlantic.
Full-text available
Predicting marine carbon sequestration in a changing climate requires mechanistic understanding of the processes controlling sinking particle flux under different climatic conditions. The recent occurrence of a warm anomaly (2014–2015) followed by an El Niño (2015–2016) in the southern sector of the California Current System presented an opportunity to analyze changes in the biological carbon pump in response to altered climate forcing. We compare primary production, mesozooplankton grazing, and carbon export from the euphotic zone during quasi-Lagrangian experiments conducted in contrasting conditions: two cruises during warm years - one during the warm anomaly in 2014 and one toward the end of El Niño 2016 – and three cruises during El Niño-neutral years. Results showed no substantial differences in the relationships between vertical carbon export and its presumed drivers (primary production, mesozooplankton grazing) between warm and neutral years. Mesozooplankton fecal pellet enumeration and phaeopigment measurements both showed that fecal pellets were the dominant contributor to export in productive upwelling regions. In more oligotrophic regions, fluxes were dominated by amorphous marine snow with negligible pigment content. We found no evidence for a significant shift in the relationship between mesozooplankton grazing rate and chlorophyll concentration. However, mass-specific grazing rates were lower at low-to-moderate chlorophyll concentrations during warm years relative to neutral years. We also detected a significant difference in the relationship between phytoplankton primary production and photosynthetically active radiation between years: at similar irradiance and nutrient concentrations, productivity decreased during the warm events. Whether these changes resulted from species composition changes remains to be determined. Overall, our results suggest that the processes driving export remain similar during different climate conditions, but that species compositional changes or other structural changes require further attention.
Full-text available
The seasonal and interannual variability of chlorophyll in the Gulf of Mexico open waters is studied using a three-dimensional coupled physical-biogeochemical model. A 5 years hindcast driven by realistic open-boundary conditions, atmospheric forcings, and freshwater discharges from rivers is performed. The use of recent in situ observations allowed an in-depth evaluation of the model nutrient and chlorophyll seasonal distributions, including the chlorophyll vertical structure. We find that different chlorophyll patterns of temporal variability coexist in the deep basin which thereby cannot be considered as a homogeneous region with respect to chlorophyll dynamics. A partitioning of the Gulf of Mexico open waters based on the winter chlorophyll concentration increase is then proposed. This partition is basically explained by the amount of nutrients injected into the euphotic layer which is highly constrained by the dynamic of the winter mixed layer. The seasonal and interannual variability appears to be affected by the variability of atmospheric fluxes and mesoscale dynamics (Loop Current eddies in particular). Finally, estimates of primary production in the deep basin are provided.
Full-text available
Functional ocean biogeography Marine ecosystems are well represented in metagenomic and transcriptomic data. These data are not routinely used to test ecosystem models that explore ocean biogeography or biogeochemistry. Coles et al. built a model in which genes for a range of functions were assigned to different suites of simulated microbes (see the Perspective by Rynearson). Communities emerged from the model with realistic biogeographical and biogeochemical profiles when compared to microbial data collected from the Amazon River plume. However, functional composition trumped the details of taxonomy, and different, coevolving community compositions emerged that provided similar biogeochemical outcomes. Science , this issue p. 1149 ; see also p. 1129
Compared to high-latitude seas, the ecological implications of climate change for top consumers in subtropical regions are poorly understood. One critical area of knowledge deficiency is the nature of food-web connections to larvae during their vulnerable time in the plankton. Bluefin tuna (BFT) are highly migratory temperate species whose early life stages are spent in ultra-oligotrophic subtropical waters. Dietary studies of BFT larvae provide evidence of prey-limited growth coupled with strong selection for specific prey types-cladocerans and poecilostomatoid copepods-whose paradoxical or poorly resolved trophic characteristics do not fit the conventional understanding of open-ocean food web structure and flows. Current knowledge consequently leaves many uncertainties in climate change effects, including the possibility that increased nitrogen fixation by Trichodesmium spp. might enhance resiliency of BFT larvae, despite a projected overall decline in system productivity. To advance understanding and future predictions, the complementary perspectives of oceanographers and fisheries researchers need to come together in studies that focus on the trophic pathways most relevant to fish larvae, the factors that drive variability in spawning regions, and their effects on larval feeding, growth and survival.
This manuscript reviews methodological differences among modified dilution assay studies used to partition phytoplankton mortality into virus‐ and grazer‐mediated fractions, and discusses their implications. A meta‐analysis is also described, based on virus‐ and grazer‐mediated mortality and instantaneous growth rates extracted from these studies. As the α value used to assess the significance of these rates was not consistent across studies, metadata was re‐analyzed (i.e., rates were re‐calculated) using three levels of significance: α = 0.05, α = 0.1, and α‐not‐considered. The average virus‐mediated mortality rate observed for each re‐analysis was 0.04 d−1, 0.10 d−1, and 0.08 d−1, respectively. Further, 80%, 74%, and 56% of virus‐mediated mortality rates were between −0.1 d−1 and 0.1 d−1. Power analysis was used to demonstrate that typical modified dilution assays lack the sensitivity to consistently detect small virus‐mediated mortality rates, and shows how dilution scheme modifications can affect power. Rates of virus‐mediated mortality were also discriminated based on salinity, depth, use of nutrient amendments, and choice of technique for estimating apparent growth rates. For all three re‐analyses, a significant positive correlation was observed between virus‐mediated mortality (d−1) and instantaneous growth rate (d−1) (Pearson r = 0.72, 0.74, 0.32, all p values < 0.01), suggesting that virally mediated mortality is intimately linked to host growth. This manuscript highlights the need for greater standardization in the analysis and presentation of information when using the modified dilution approach for estimating phytoplankton mortality, and provides recommendations for the application of future assays.
Information on early life history of economical important fisheries stocks are required to accurately estimate their population status. This study investigated blackfin tuna (Thunnus atlanticus) larvae distribution over six summers (2007–2011, 2015) in the northern Gulf of Mexico. Blackfin tuna were commonly observed and widely distributed in surface waters with frequency of occurrence ranging from 48% (2008) to 92% (2011). Inter-annual variability in density was observed with highest mean density recorded in 2009 (17.2 larvae 1000 m⁻³) and lowest mean density in 2015 (2.2 larvae 1000 m⁻³). Density also varied between months with higher overall mean density observed in July (9.2 larvae 1000 m⁻³) compared to June (4.3 larvae 1000 m⁻³). Generalized additive models (GAMs) based on presence/absence and density of blackfin tuna larvae determined that this species was present in areas of intermediate salinity (31–36) and higher sea surface temperature (SST > 29 °C). Blackfin tuna larvae were also strongly associated with convergent zones near the Loop Current and anticyclonic eddies. Environmental conditions deemed to be favorable from GAMs (salinity, SST and sea surface height) were combined with environmental data in 2011 and 2015 to predict the suitable habitat of blackfin tuna larvae from the outer continental shelf into oceanic waters (areas ≥100 m isobath). The amount of highly suitable habitat ( >10 larvae 1000 m⁻³) in 2011 and 2015 varied between months (June 6%, July 51%); however, blackfin tuna larvae were predicted to occur at similar locations in surface waters along the continental slope and at the margin of the Loop Current. Overall, the results highlighted the importance of mesoscale features and oceanographic conditions on the distribution and abundance of blackfin tuna larvae.