Art 1

Progress in Oceanography 134 (2015) 54–76
Contents lists available at ScienceDirect
Progress in Oceanography
journal homepage: www.elsevier.com/locate/pocean
Natural variability of surface oceanographic conditions in the offshore

Gulf of Mexico
Frank E. Muller-Karger a,⇑, Joseph P. Smith b, Sandra Werner b, Robert Chen a, Mitchell Roffer c,
Yanyun Liu d,g, Barbara Muhling d,e, David Lindo-Atichati d, John Lamkin e, Sergio Cerdeira-Estrada f,
David B. Enfield d
a
College of Marine Science, University of South Florida, St. Petersburg, FL, USA
b
ExxonMobil Upstream Research Company, Houston, TX, USA
c
Roffer’s Ocean Fish Forecasting Service (ROFFSTM), Melbourne, FL 32904, USA
d
University of Miami, Cooperative Institute for Marine and Atmospheric Studies, Miami, FL 33149, USA
e
National Oceanic and Atmospheric Administration, Southeast Fisheries Science Center, Miami, FL 33149, USA
f
Comisión Nacional para el Conocimiento y Uso de la Biodiversidad (CONABIO), Mexico City, DF, Mexico
g
Atlantic Oceanographic and Meteorological Laboratory, NOAA, Miami, FL, USA
a r t i c l e i n f o a b s t r a c t
Article history: This work characterizes patterns of temporal variability in surface waters of the central Gulf of Mexico.
Received 9 August 2014 We examine remote-sensing based observations of sea surface temperature (SST), wind speed, sea sur-
Received in revised form 25 October 2014 face height anomaly (SSHA), chlorophyll-a concentration (Chl-a) and Net Primary Production (NPP), along
Accepted 8 December 2014
with model predictions of mixed layer depth (MLD), to determine seasonal changes and long-term trends
Available online 29 December 2014
in the central Gulf of Mexico between the early 1980s and 2012. Specifically, we examine variability in
four quadrants of the Gulf of Mexico (water depth >1000 m). All variables show strong seasonality.
Chl-a and NPP show positive anomalies in response to short-term increases in wind speed and to cold
temperature events. The depth of the mixed layer (MLD) directly and significantly affects primary pro-
ductivity throughout the region. This relationship is sufficiently robust to enable real-time estimates of
MLD based on satellite-based estimates of NPP. Over the past 15–20 years, SST, wind speed, and SSHA
show a statistically significant, gradual increase. However, Chl-a and NPP show no significant trends over
this period. There has also been no trend in the MLD in the Gulf of Mexico interior. The positive long-term
trend in wind speed and SST anomalies is consistent with the warming phase of the Atlantic Multidecadal
Oscillation (AMO) that started in the mid-90s. This also coincides with a negative trend in the El Niño/
Southern Oscillation Multivariate ENSO Index (MEI) related to an increase in the frequency of cooler
ENSO events since 1999–2000. The results suggest that over decadal scales, increasing temperature, wind
speed, and mesoscale ocean activity have offsetting effects on the MLD. The lack of a trend in MLD anom-
alies over the past 20 years explains the lack of long-term changes in chlorophyll concentration and pro-
ductivity over this period in the Gulf. Understanding the background of seasonal and long-term
variability in these ocean characteristics is important to interpret changes in ocean health due to episodic
natural and anthropogenic events and long term climate changes or development activities. With this
analysis we provide a baseline against which such changes can be measured.
Ó 2014 The Authors. Published by Elsevier Ltd. Open access under CC BY-NC-ND license.
Introduction thruway for the transport of heat, salt and nutrients, and biological
material from the Caribbean Sea to the North Atlantic. Thus, it
The Gulf of Mexico (Fig. 1) is an area of great economic impor- plays a significant role in defining the weather and climate of Cen-
tance for the United States, Mexico, and Cuba (Adams et al., 2004). tral America, the United States, and the Caribbean Sea. It represents
It forms part of the circulation of the Atlantic Ocean, serving as a important habitat for pelagic fish, marine mammals, and other
organisms. A unique geography, rich biodiversity, and high produc-
tivity are the foundations for major fisheries and tourism indus-
tries in the Gulf of Mexico. This is also an area that has led the
⇑ Corresponding author. Address: College of Marine Science, University of South
world in extensive oil and gas development in deep ocean waters
Florida, 140 7[th] Avenue South, St. Petersburg, FL, USA. Tel.: +1 727 553 3335.
E-mail address: carib@usf.edu (F.E. Muller-Karger). (>300 m) since the 1970s.
http://dx.doi.org/10.1016/j.pocean.2014.12.007
0079-6611/Ó 2014 The Authors. Published by Elsevier Ltd. Open access under CC BY-NC-ND license.
F.E. Muller-Karger et al. / Progress in Oceanography 134 (2015) 54–76 55
Fig. 1. The Gulf of Mexico. Image shows the four regions of interest (ROI) used to develop the climatological analyses of wind, sea surface temperature, and chlorophyll-a
concentration. ROIs referenced in the text are labeled in this figure as 1-NE, 2-NW, 3-SE, and 4-SW. The locations of the NOAA NDBC buoys used for wind data are marked by
filled squares; each buoy location is labeled with the NDBC buoy number.
Over the past 60 years we have developed substantial knowl- away from the more direct influence of winds and atmospheric
edge about the hydrography, circulation, biogeochemical, and bio- temperatures associated with adjacent land, which strongly affect
logical processes of surface waters of the Gulf of Mexico (e.g. coastal and shelf waters of the Gulf (Muller-Karger, 2000; Zavala-
Capurro and Reid, 1972; Muller-Karger et al., 1991; Felder and Hidalgo et al., 2003; Morey et al., 2005; Weisberg et al., 1996,
Camp, 2012; and numerous other publications). This knowledge 2005). Surface waters of the Gulf of Mexico cover approximately
is based on observations collected by ships, moored instruments, 1.6 106 km2, with waters deeper than the continental shelf cov-
drifting buoys, satellite-based sensors, and numerical simulations. ering about 8.6 105 km2. Satellite sensors facilitate the observa-
In this paper we present an updated synthesis of the major charac- tion of such large areas. We examined the variability of SST,
teristics of oceanic surface waters in the central Gulf of Mexico. wind speed, sea surface height anomaly (SSHA), mixed layer depth
Our study was guided by the objective of testing the hypothesis (MLD), chlorophyll-a concentration, and primary productivity in
that the surface waters of the interior Gulf of Mexico, i.e. those sea- these offshore Gulf of Mexico waters. We sought to quantify ocean-
ward of the shelf break, have experienced significant unidirectional ographic change over seasonal to decadal timescales and put these
change over the past 30 years in a number of oceanographic variations in the context of decadal-scale change. The need for such
parameters. Detecting and quantifying such change in large oce- information is recognized, especially after the Deepwater Horizon
anic ecosystems is fundamental to define proper resource use oil spill (e.g. National Ocean Service, 2011).
management strategies (Muller-Karger et al., 2014). Our question
arose from general interest on how the Gulf of Mexico is behaving Background
in light of the average +0.13 °C per decade increase in sea surface
temperature observed over the global ocean since 1979 There is a large body of scientific literature describing the gen-
(Trenberth et al., 2007), or the much higher rates of +0.5 °C per dec- eral circulation of surface waters in the Gulf of Mexico (Nowlin
ade observed in the past 20 years over the central Caribbean Sea et al., 1968; Behringer et al., 1977; Vukovich, 1988; Fratantoni
and the tropical western north Atlantic Ocean (Chollett et al., et al., 1998; Lindo-Atichati et al., 2013, and references therein).
2012). Muhling et al. (2012) found that sea surface temperature The Gulf of Mexico forms part of the western boundary current
(SST) in the northern Gulf of Mexico rose about 0.5 °C between system of the North Atlantic. Clear and warm surface Caribbean
1985 and 2008. They detected a concurrent increase in pelagic fish Sea water enters the basin via the Yucatan Current. This water
larvae over the outer continental shelf, with warmer years showing can penetrate as far north as 29°N, reaching the vicinity of the Mis-
higher abundances of larvae. Muhling et al. (2012) concluded that sissippi River delta in the northern Gulf of Mexico. This current
long-term changes in physical habitat are having important eco- transports water volumes between about 24 Sverdrup (Sv)
logical implications in the deep Gulf of Mexico. (Sheinbaum et al., 2002) and 32 Sv (Morrison and Nowlin, 1977;
Here we focused specifically on key surface ocean properties of Baringer and Larsen, 2001; others). Many studies have focused
the interior of the Gulf of Mexico where the seafloor is deeper than on possible links between the volume transport through Yucatan
1000 m. We wanted to understand long-term change in waters Channel, the extent of penetration of the Loop Current into the Gulf
56 F.E. Muller-Karger et al. / Progress in Oceanography 134 (2015) 54–76
of Mexico, and the processes of ring shedding (Maul et al., 1985; 1983; Cerdeira-Estrada et al., 2005; Chang and Oey, 2010). Heat
Candela et al., 2003; Bunge et al., 2002; Oey, 1996; Ezer et al., loss in winter is accelerated by storms and cold-air fronts that
2003). Oscillations in the Cayman Sea and the Gulf of Mexico were stimulate Ekman pumping and wind-driven and convective mixing
hypothesized to be connected with time-dependent growth of the (Muller-Karger et al., 1991; Melo Gonzalez et al., 2000; Villanueva
Loop Current and its penetration into the northern areas of the Gulf et al., 2010). Muller-Karger et al. (1991) used SST data from the
(Maul, 1978). As the current extends to the north, it makes a loop NOAA Advanced Very High Resolution Radiometer (AVHRR) and
by turning east and south, and carries water out of the Gulf via the the Comprehensive Ocean–Atmosphere Data Set (COADS) to docu-
Straits of Florida (Vukovich et al., 1979; Hurlburt and Thompson, ment the seasonal cycle of SST in the interior of the Gulf and its
1980). The Loop Current penetration displaces an equivalent vol- relation to the mixed layer depth. Highest SST of 29–30 °C occur
ume of Gulf water which flows back into the Caribbean Sea approx- throughout the interior of the Gulf between July and September.
imately below 800 m and in shallower counter-currents found SST minima occur in February–March, with 22–24 °C observed
typically on the eastern side of the Yucatan Channel (Bunge in the western half of the Gulf, and 24–26 °C in eastern half
et al., 2002; Candela et al. 2003; Rivas et al. 2005; Chérubin et al. where the Loop Current has a direct influence. The density changes
2005). Ocean color satellite imagery shows phytoplankton blooms associated with temperature and the changes in wind forcing lead
from the Yucatan Peninsula are often entrained into the Yucatan/ to a seasonal cycle in mixed layer depth (MLD), with shallow
Loop Current system and transported north and east into the Gulf MLD 20 m during boreal summers and 125 m in winter
of Mexico interior (Muller-Karger et al., 1991). Similarly, the Loop (Muller-Karger et al., 1991).
Current often entrains water from the Mississippi River along its Even today there are very limited field observations available to
northern edge and may carry it out of the Gulf of Mexico via the characterize how large-scale physical forces have affected the basic
Straits of Florida (Muller-Karger, 1993; Ortner et al., 1995; Del biogeochemical characteristics of the interior of the Gulf of Mexico
Castillo et al., 2000, 2001; Hu et al., 2005). The Loop Current shows over time. Such inferences have been facilitated by satellite obser-
instabilities around its edge, many of which form cyclonic eddies vations. Muller-Karger et al. (1991) examined the spatial and tem-
that may grow as they propagate around its periphery. As these poral changes in phytoplankton concentrations over the Gulf of
eddies grow, they likely play a role in the process of separation Mexico region using synoptic ocean color satellite imagery col-
of a section of the Loop Current into a large (200–300 km diameter) lected with the NASA Coastal Zone Color Scanner (CZCS; 1978–
anticyclonic eddy (Fratantoni et al., 1998). 1986). They found a seasonal variation in phytoplankton concen-
The process of Loop Current extension and anticyclonic eddy tration seaward of the shelf, with lowest values in May–July
separation is the result of a number of interacting seasonal and sto- (<0.06 mg m 3) and high values in December–February
chastic processes (Nowlin et al., 2000; Zavala-Hidalgo et al., 2006). (>0.2 mg m 3). Winter fronts and storms are particularly frequent
In rare years no separation is observed, but separations can occur during El Niño-Southern Oscillation years (Melo Gonzalez et al.,
even up to three times per year. The period between eddies varies 2000; Kennedy et al., 2007). These storms lead to increased mixing
between 0.5 and 18.5 months (Vukovich 2007, 2012; Sturges and and higher phytoplankton concentrations in the deep Gulf of Mex-
Leben, 2000; Leben 2005). The time series of satellite altimetry ini- ico during El Niño years (Melo Gonzalez et al., 2000).
tiated in 1992 shows anticyclonic eddies shed more frequently Martínez-López and Zavala-Hidalgo (2009) derived a synthetic
between about June and September, but eddies have been shed 12-month climatology of chlorophyll-a concentrations using esti-
any month of the year (Alvera-Azcárate et al., 2009; Lindo- mates derived by NASA using the Sea-viewing Wide-Field-of-view
Atichati et al., 2013; Cardona and Bracco, 2014). Similar inferences Sensor (SeaWiFS) for the period 1997–2007. They found the highest
have been made from simulations (Chang and Oey, 2010, 2012, variations in chlorophyll-a concentration on the shelf and along the
2013a,b; Nedbor-Gross et al., 2014). shelf-break. Using the climatology, they identified complex tempo-
In the eastern Gulf of Mexico, the Loop Current and these large ral patterns of wind-driven export of shelf materials to the interior of
eddies can interact with the continental shelf and lead to enhanced the Gulf off the Mississippi River Delta, off the Louisiana-Texas shelf,
coastal upwelling throughout the region (Muller-Karger, 2000; off Veracruz, and in the southern Bay of Campeche. Salmerón-García
Weisberg et al., 2004). The large anticyclonic Loop Current eddies et al. (2011) also used long-term monthly mean SeaWiFS chloro-
typically drift toward the west or west-northwest and reach the phyll-a concentration estimates to derive a thematic climatological
Texas–Mexico coastal zone after 2–4 months (Muller-Karger (1998–2008) classification of Gulf of Mexico biogeographic regions.
et al., 1991; Lindo-Atichati et al., 2013 and references therein). Callejas-Jimenez et al. (2012) conducted a similar spatial climato-
Immediately after shedding an eddy, the Loop Current retracts to logical (2002–2007) classification but combining SST, chlorophyll-
a southern position where waters flow more directly between a, and normalized water-leaving radiance observations derived by
Yucatan Channel and the Straits of Florida. There is much debate NASA using the Moderate Resolution Imaging Spectrometer (MODIS
on the physical oceanographic processes at play here. Le Hénaff on the Aqua satellite). These studies all reached the conclusion that
et al. (2012) suggested that blockage is effected by one or more continental shelf waters (i.e. those where the bottom is shallower
successive cyclones north of the Loop Current boundary. Mildner than about 200 m) host upwards of a dozen different biogeographic
et al. (2013) suggested that the recently-shed anticyclonic eddy regions distributed around the periphery of the Gulf. Waters sea-
blocks the flow of the Yucatan Current water toward the north. ward of the continental shelf were generally classified as a single
This remains a ‘‘chicken and the egg’’ problem that will require biogeographic region, with the caveat that the northeastern and
study outside the scope of this paper. northwestern corners of the Gulf experience seasonal offshore
Many of the studies mentioned above have demonstrated that advection of coastal waters (Muller-Karger et al., 1991; Muller-
the bulk properties of the water masses of the upper Gulf of Mexico Karger, 2000; Biggs and Muller-Karger, 1993; Biggs et al., 2005;
are determined by interaction of the atmosphere and the ocean Martínez-López and Zavala-Hidalgo, 2009).
over seasonal cycles, by mixing of Loop Current water and large The analysis we present here updates these studies for surface
anticyclonic and cyclonic eddies, and by upwelling and mixing of waters over the deep Gulf of Mexico. The study is based on time
waters along the margins of the basin (Herring, 2010; Vidal et al., series of synoptic satellite observations collected from the 1980s
1994, and references therein). Sea surface temperature (SST) in through 2012, with more limited observations collected with the
the interior of the Gulf of Mexico undergoes a marked seasonal Coastal Zone Color Scanner starting in 1978. The observations
cycle. Waters gain heat through insolation between April and allow the characterization of the variability in physical parameters
August, but lose heat between September and March (Etter, that affect the biological production of surface ocean waters.
More specifically, among the physical variables we examined Time series of NPP and of Chl-a derived from satellites thus pro-
were SST, sea surface height (SSH), wind speed, and MLD. SST is vide proxy information on trophic dynamics of surface waters of
an index of the turbulent and thermodynamic heat flux balance the ocean that is helpful to study the impact of variations in
of the surface ocean. It plays a role in determining the rate of met- near-surface water column thermal structure and mixing.
abolic processes within organisms, and is used by biota as a cue for About two decades of synoptic observations of all of these
migratory, reproductive, and feeding behavior. Satellite SST obser- parameters are now available. The phytoplankton pigment concen-
vations provide the most comprehensive, long-term records of tration time series derived from SeaWiFS (1997–2010) and MODIS-
daily physical variability in the world’s ocean. Wind speed contrib- Aqua (2002–2013) provide excellent data with which we can re-
utes to the vertical pumping of nutrients and mixing of biogeo- examine the inferences made about biological variability based
chemical properties in the upper layers of the ocean. Wind also on the CZCS observations collected between 1978 and 1986
contributes to the heat flux balance and gas exchange between (Muller-Karger et al., 1991; Melo Gonzalez et al., 2000). For this
the ocean’s surface and the atmosphere. SSH is strongly related purpose we also used a new version of the CZCS dataset, repro-
to circulation patterns and to the internal thermal and salinity cessed by NASA in 2011 using bio-optical algorithms that are con-
structure of the ocean (Rio and Hernandez, 2004). In our region, sistent with those used by SeaWiFS and MODIS.
a higher SSH than normal can be caused by an anticyclonic meso- We also compared the variability in these observations with
scale circulation feature, but it can also indicate the presence of a common indices of climate variability, in particular the Atlantic
deep layer of warmer than average water. Depending on many fac- Multidecadal Oscillation (AMO; Enfield et al., 2001) and the El
tors, such as the vertical stratification and the dynamic processes Niño/Southern Oscillation Multivariate ENSO Index (ENSO MEI;
involved, the relationship between isotherms and the SSH can be Wolter and Timlin; 2011). The AMO changed from a warm to cool
estimated from altimeter-derived SSHA, in combination with in phase in the mid-60s and from a cool to warm phase in the mid-
situ and climatological hydrographic observations. In general, vari- 90s (Enfield and Cid-Serrano 2006, 2010). The Gulf of Mexico eco-
ations in the depth of the main thermocline can be associated with system seems to have shifted in response to this change in the
variations in the SSHA field (Willis et al. 2004; Shay et al., 2000). phase of the AMO (Karnauskas et al., 2013). Simulations under var-
The MLD defines the portion of the water column immediately ious future climate scenarios also suggest that the Loop Current
below the surface within which physical and biogeochemical vari- (LC) in the Gulf of Mexico may slow down as part of a deceleration
ables are more or less homogeneous. The MLD is the result of the of the Atlantic Meridional Overturning Circulation (AMOC), one of
interaction between surface and upper ocean processes of surface the major mechanisms affecting variability in the AMO (Liu et al.,
heating and cooling, wind and convective mixing, molecular diffu- 2012).
sion and horizontal advection. Thus, the MLD has implications for With this study we provide a baseline against which such
climate and weather in terms of heat and material exchanges changes can be measured in the Gulf of Mexico. We revisit the con-
between the atmosphere and the deep ocean. It also has important clusions of Muller-Karger et al. (1991) that the seasonal phyto-
implications for biota, in terms of light limitation and nutrient plankton biomass is defined by variations in the mixed layer, and
availability (Sverdrup, 1953; Ryther and Menzel, 1960; Menzel examine how interannual variability in environmental parameters
and Ryther, 1961). may affect the primary productivity of Gulf waters over decadal
Among biological parameters that serve as an index to the bio- scales.
logical state of the Gulf of Mexico we examined Chlorophyll-a
(Chl-a) and Net Primary Production (NPP). Chl-a is an index of
the standing stock of phytoplankton in ocean waters, typically Methods
expressed in units of biomass per unit volume of water. Chl-a
also serves as an indicator of vertical mixing of nutrients in the To test the hypothesis of whether the oceanography of surface
upper ocean and as a tracer of the horizontal dispersal of upwell- waters of the Gulf of Mexico has changed over the past 20+ years,
ing or river plumes. Chl-a is typically a small fraction of the par- we developed decadal-scale time series of various surface ocean
ticulate organic carbon (POC) in the water (i.e. a few percent). parameters for the Gulf of Mexico based on satellite observations.
Carbon to Chl-a ratios vary widely (C:Chl-a can range from < 10 We divided the Gulf into four quadrants (Fig. 1). Regions of Interest
to >300 mg C mg Chl-a 1), with lower values typical of diatoms (ROI) were defined as areas between the 1000 m isobath and a
and/or nutrient replete conditions (Yoder, 1979; Laws and point located at 27.78°N, 89.86°W in the central Gulf of Mexico.
Bannister, 1980; Falkowski et al., 1985; Li et al., 2010; and many This point was chosen arbitrarily to divide the Gulf roughly into
others). Higher values (>200 mg C mg Chl-a 1) are more typical of east, west, north and south. The northern quadrants coincide
non-diatom and/or nutrient limiting conditions. The C:Chl-a ratio roughly with the U.S. Exclusive Economic Zone (EEZ). The eastern
also tends to decrease with depth toward the deep Chl-a maxi- quadrants represent the area which is influenced by the Loop Cur-
mum, below which C:Chl-a may again increase with depth. There rent and where the anticyclonic eddy shedding process occurs. We
still are very few systematic time series measurements of this examine variability and trends in wind intensity, sea surface tem-
parameter in the world’s ocean. We did not find any published perature (SST), sea surface height anomaly (SSHA), phytoplankton
time series of C:Chl-a for the interior of the Gulf of Mexico in pigment concentration (chlorophyll-a or Chl-a) and Net Primary
waters deeper than the continental shelf edge. In our discussion Production (NPP) derived from satellite observations for each of
below we explore the implications of variations in the C:Chl-a these quadrants. Weekly, monthly and annual arithmetic means,
with season as related to Chl-a biomass observations from satel- and the corresponding long-term means (i.e. ‘‘climatologies’’) were
lite-based sensors. computed for each parameter.
NPP is a measure of the rate of carbon fixation by phytoplank- Time series of anomalies of wind speed, SST, SSHA and Chl-a con-
ton (photosynthesis). It is typically expressed as the biomass (in centration were obtained by subtracting the long-term monthly
units of weight of carbon) under a square meter of the ocean per mean (climatology) from the monthly field for that variable. Anom-
unit time. NPP is a complex function of the physiology of phyto- alies thus represent de-seasoned decadal-scale time series of the
plankton, growth rate, C:Chl-a ratios, temperature, nutrient, and corresponding observables. They were used to examine trends using
sunlight and nutrient history and availability (Cloern et al., 1995; a test for significance of the difference between the slope of the trend
Li et al., 2010). derived by least-squares linear regression and a slope of zero. Using
anomalies helped to mitigate the effect of autocorrelation in exam- – Buoy 42003 (26.044°N, 85.612°W; east-central Gulf; 3283 m;
ining the characteristics of these time series. 1976–2012 inclusive).
Wind (1987–2011) We derived long-term monthly average scalar wind speed from
these buoys to compare with the CCMP wind ‘climatology’ men-
We used monthly averages (1987–2011) of the Cross-Cali- tioned above (Fig. 2). The buoy data were also used to validate
brated, Multi-Platform Ocean Surface Wind Velocity Product any inferences of long-term trends in wind speed. The buoys were
(CCMP; Atlas et al., 2011) available from NASA (ftp://podaac-ftp. useful to evaluate the representativeness of the CCMP wind statis-
jpl.nasa.gov/allData/ccmp/L3.0/docs/ccmp_users_guide.pdf). This tics for different ROI.
ocean surface (10 m) wind product, mapped to a 0.25 0.25°
degree cylindrical grid, incorporates cross-calibrated satellite Sea surface temperature (1982–2013)
winds derived from SSM/I, SSMIS, AMSR-E, TRMM TMI, QuikSCAT,
SeaWinds, WindSat, and other satellite instruments. The product is SST was derived from infrared (IR) observations collected by a
assumed to avoid the diurnal cycle bias that may be expected in number of different satellites. Data were processed at different
observations from single polar-orbiting wind sensors (Tang et al., spatial and temporal resolutions for various tests conducted for
2014). We specifically used scalar wind speed and derived long- our study. Specifically, a time series of daily SST observations span-
term monthly average winds (i.e. a ‘climatology’) to examine ning 1982–2012 (inclusive) was extracted from the Advanced Very
anomalies and long-term trends in the winds experienced in the High Resolution Radiometer (AVHRR) Pathfinder Version 5.2
region. (PFV5.2). The nominal spatial resolution of these data is
To evaluate the quality of the satellite-derived ocean wind data, 4 4 km2 per pixel. The data were obtained from the US National
wind observations were extracted from three oceanographic buoys Oceanographic Data Center (NODC) and the Group for High Resolu-
maintained in the central Gulf of Mexico since the mid-1970s by tion Sea Surface Temperature (GHRSST) (http://pathfinder.nodc.
the National Oceanic and Atmospheric Administration’s National noaa.gov; Casey et al., 2010). The Pathfinder v5.2 dataset has a
Data Buoy Center (NOAA NDBC). Specifically, we selected the fol- gap, from October 2, 1994 to January 17, 1995. This gap was filled
lowing NDBC buoys, identified by their number (Fig. 1): for our Gulf of Mexico time series with Local Area Coverage (LAC)
AVHRR data collected and processed at the University of South
– Buoy 42001 (25.888°N, 89.658°W; central Gulf; depth of Florida (USF). The daily AVHRR LAC data cover the period August
3365 m; 1975–2012 inclusive). 1993 to 2013. These images were processed at a spatial resolution
– Buoy 42002 (25.790°N, 93.666°W; west-central Gulf; 3566 m; of approximately 1 1 km2 per pixel. Likewise, daily data from the
1973–2012 inclusive; note that this buoy was previously posi- MODIS Aqua sensor were processed to estimate SST using the
tioned at 25.167°N, 94.417°W, about 60 miles to the SW. This 11 lm band at approximately 1 1 km2 per pixel.
change in position is small and does not lead to a change in We computed long-term weekly and monthly mean and stan-
results). dard deviation (SD) time series for these datasets by binning all
Buoy 42002 Buoy 42003
Buoy 42002 Buoy 42001
Fig. 2. Wind speed monthly climatology for the four ROI’s defined for the interior Gulf of Mexico (Fig. 1). Regions of Interest: Northeast (upper right), Northwest (upper left),
Southeast (lower right), and Southwest (lower left) quadrants (Fig. 1). Solid curves show the monthly mean climatology (1987–2011) derived from the Cross-Calibrated,
Multi-Platform Ocean Surface Wind Velocity Product (CCM). The stippled curves on either side show the mean ± one standard deviation. The standard deviation was derived
on a per pixel basis while constructing the long-term climatological means, and then was averaged over each ROI for each climatological month. The broken line overlaid on
each climatological mean wind speed curve is the long-term monthly wind speed means derived from NOAA buoys (Buoy 42001: 1975–2012 inclusive; Buoy 42002: 1973–
2012 inclusive; Buoy 42003: 1976–2012 inclusive).
daily night-time only data from each sensor, and for each month fields have a spatial resolution of 0.25 0.25° and temporal reso-
across years for each pixel. The Pathfinder v5.2 SST climatology lution of 1 week.
thus is represented by 52 synthetic weekly means and SD, or 12 To examine long-term variability in SSHA in the region, we
values for the monthly mean and SD, constructed over the period computed long-term monthly mean and standard deviation (SD)
1982–2010 (inclusive). The October 1994–January 1995 gap in of the AVISO datasets (Fig. 4). These climatologies were derived
Pathfinder v5.2 data mentioned above was not filled for purposes by binning all weekly data. The AVISO SSHA climatology is repre-
of computing the climatology. The MODIS and AVHRR SST monthly sented by 12 values for the monthly mean and standard deviation
climatology (mean and SD) were computed for the period 2003– constructed over the period October 1992–December 2012 (inclu-
2010 (inclusive). The choice of years was chosen since it is the per- sive). We also examined the decadal-scale variability and trends in
iod of overlap with SeaWiFS observations. the de-seasoned SSHA (i.e. the monthly SSHA anomaly) as an indi-
The Pathfinder v5.2 SST time series for the Gulf of Mexico was cator of mean sea surface elevation change and to assess ocean
compared against the USF AVHRR and MODIS Aqua data. These mesoscale activity.
long-term, night-time SST climatologies were similar in each of Because we analyzed climatologies and time series of anoma-
the four Gulf of Mexico quadrants (Fig. 3). The similarity provides lies, we did not add the mean sea surface elevation to the SSHA
confidence in the accuracy of the data and in long-term patterns fields. The AVISO sea surface height anomaly (SSHA) fields were
observed with them. In the past, we have found that the various computed with respect to the 1993–1999 mean from direct altim-
satellite data underestimate in situ observations slightly, in the etry observations.
order of about 0.5 °C (Hu et al., 2009). In the present study we
examined SST anomalies and long-term trends, and therefore Mixed layer depth
ignored this small bias.
We sought to derive a time series of monthly mean mixed layer
Sea surface height anomaly (1992 – 2012) depth estimates for the four offshore Gulf of Mexico quadrants
using all in situ data holdings of the NOAA National Oceanographic
The sea surface height anomaly (SSHA) is the difference Data Center (NODC). After downloading all ship cast, XBT, drifting
between the best estimate of the sea surface height and a mean buoy, and glider data available at NODC, we found that the data
sea surface derived from long-term observations from satellite still are insufficient to derive a reasonable time series to cover
altimeters. This ‘anomaly’ thus preserves seasonal signals. We the period of our study. We therefore did not use these data in this
derived a time series of monthly SSHA estimates for the four Gulf analysis. As an alternative to field measurements for assessment of
of Mexico quadrants using satellite altimetry observations gridded variability in the mixed layer, we examined daily output fields
by the French CNES AVISO office (Archiving, Validation and Inter- (1992–2012 inclusive) of the ECCO2 model (ECCO2 is an acronym
pretation of Satellite Oceanographic data; Le Traon et al., 1998). for Estimating the Circulation and Climate of the Ocean, Phase II;
The altimetric observations used for our time series correspond results available at: http://ecco2.jpl.nasa.gov/products/). ECCO2 is
to Jason-1, TOPEX/Poseidon, The European Remote Sensing (ERS) Phase II of the High-Resolution Global-Ocean and Sea-Ice Data Syn-
Satellites 1 and 2, the Environmental Satellite (ENVISAT), and the thesis, sponsored by the NASA Modeling, Analysis, and Prediction
Geodetic Satellite (GEOSAT) and the GEOSAT Follow-On (GFO), (MAP) program (Menemenlis et al., 2005a,b, 2008; Wunsch et al.,
with data starting in October 1992. These interpolated gridded 2009). The underlying model is the Massachusetts Institute of
Fig. 3. SST monthly climatology for the four ROI’s defined for the interior Gulf of Mexico (Fig. 1) from the night-time NOAA AVHRR Pathfinder v5.2 (solid line; 1982–2010)
and the NASA night-time MODIS 11 lm SST product (broken line; 2003–2010). The stippled curves on either side show the mean ± one standard deviation; the standard
deviation were derived from the daily AVHRR Pathfinder v 5.2 SST data, on a per pixel basis while constructing the long-term climatological means, and then these were
averaged over each ROI for each climatological month.
Fig. 4. SSHA monthly climatology for the four ROI’s defined for the interior Gulf of Mexico (Fig. 1) from weekly averaged AVISO fields (solid line; 1993–2012 inclusive). The
stippled curves on either side show the mean ± one standard deviation around the long-term monthly SSHA estimates for each region.
Technology general circulation model (MITgcm; Marshall et al., with detritus and colored dissolved organic matter or CDOM con-
1997a,b). centration (Morel and Prieur, 1977). Under such conditions, ocean
ECCO2 is initialized from the World Ocean Database (Conkright color data are an appropriate tool for long term characterization of
et al., 2002) and forced by surface fluxes (wind stress, heat, and chlorophyll-a variability. Time series of chlorophyll-a concentra-
freshwater) from the NCEP meteorological synthesis (Kalnay tion images of the Gulf of Mexico were derived from data collected
et al., 1996; Kistler et al., 2001). The spatial resolution is 1/4° with by the Coastal Zone Color Scanner (CZCS), the Sea-viewing Wide-
50 vertical levels ranging in thickness from 10 m near the surface Field-of-view Sensor (SeaWiFS), and the Moderate Resolution
to approximately 450 m near the bottom in the deep ocean. The Imaging Spectrometer (MODIS on NASA’s Aqua satellite). The CZCS
NCEP forcing fields are produced on a 6-hourly basis. Daily average observations span October 1978 through May 1986, when the sen-
mixed layer depth (MLD) was estimated from the ECCO2 model as sor stopped working. The SeaWiFS measurements span January
the depth at which the potential density relative to the surface is 1998 through November 2010. SeaWiFS stopped working in
larger than surface density by using the DeltaRho = 0.8°C ⁄Alpha December 2010. We merged the MODIS and SeaWiFS data series
criterion, where Alpha is the thermal expansion coefficient at the for our long-term analyses of chlorophyll concentrations, with
surface (Kara et al., 2000). Monthly climatologies (Fig. 5) were SeaWiFS data spanning 1998–2002 (inclusive). MODIS data
derived from the time series of daily estimates. spanned January 2003–January 2013.
ECCO2 model results in the Gulf of Mexico seem to be realistic Ocean color data from the CZCS, SeaWiFS, and MODIS Aqua
and consistent with observations and current understanding of satellite sensors were obtained from the NASA Goddard Space
physical oceanographic processes in the region. This includes, for Flight Center. We used the updated CZCS chlorophyll-a data
example, the timing and location of shedding of anticyclonic eddies (2011 reprocessing by NASA), SeaWiFS (2010 reprocessing), and
from the Loop Current, vertical flow speeds observed within anti- MODIS (2013 reprocessing). The 2011 CZCS reprocessing used cal-
cyclonic eddies, and computed Lagrangian Coherent Structures ibration, atmospheric correction, and band-ratio chlorophyll-a
(Lipinski and Mohseni, 2014). Tulloch et al. (2011) also concluded concentration algorithms similar to those used by SeaWiFS and
that ECCO2 provided realistic simulations of the Loop Current MODIS to address the problems listed by Gregg et al. (2002). All
eddy-shedding process, with the anticyclonic eddies subsequently products followed the latest implementation of the atmospheric
moving westward in the Gulf. They found that the ECCO2 solutions correction based on Gordon and Wang (1994), Gordon (1997),
are in agreement with the simulations of the National Center for and Ding and Gordon (1995). Chlorophyll-a concentration from
Atmospheric Research (NCAR) Climate Community System Model CZCS, SeaWiFS and MODIS was estimated using the NASA OC4
(CCSM; Maltrud et al., 2010). and OC3 band ratio algorithms (O’Reilly et al., 2000). We used
the global 4 4 km2 resolution gridded CZCS and MODIS chloro-
Chlorophyll-a concentration (mg m 3) and Net Primary Production phyll-a fields and the 9 9 km2 resolution global SeaWiFS prod-
(NPP; mg C m 2 d 1) (1978–2013) ucts for the basin-scale analyses.
Gregg et al. (2002) list some of the deficiencies of the CZCS rel-
In oligotrophic and mesotrophic waters, such as most of the ative to the SeaWiFS and MODIS sensors. They analyzed the global
deep Gulf of Mexico, phytoplankton concentration will co-vary CZCS data using a number of improvements relative to the original
Fig. 5. Mixed layer depth (MLD in meters) monthly climatology derived from daily average ECCO2 simulated MLD fields (1992–2012 inclusive). MLD was derived using the
DeltaRho = 0.8°C ⁄Alpha criterion, where Alpha is the thermal expansion coefficient at the surface criterion relative to surface temperature. The stippled curves on either side
show the mean ± one standard deviation around the long-term monthly MLD estimates for each region.
CZCS algorithms, including improved calibration and atmospheric CZCS data, however, are useful to compare general temporal and
correction algorithms. They also applied a 3-band maximum ratio spatial patterns observed in the 1980s with those seen with the
bio-optical algorithm (OC3) intended to estimate chlorophyll-a SeaWiFS and MODIS. This includes, specifically for this study, gen-
concentration, as opposed to pigment concentration (the sum of eral seasonal variability in pigment concentration, changes in spa-
chlorophyll and other pigments). They found that the original CZCS tial patterns of ocean color associated with the edges of the Loop
data generally underestimated chlorophyll concentration observa- Current and eddies, as well as corroborating variability in the sea-
tions archived by the NOAA National Oceanographic Data Center sonal dispersal patterns of Mississippi river water in the Northeast-
(NODC) globally by 8–35% (Gregg and Conkright, 2001). On regio- ern Gulf of Mexico.
nal and seasonal scales, larger underestimates were common (20– Monthly Net Primary Production (NPP) based on the Vertically
40%, and occasionally the differences exceeded 100%). The repro- Generalized Production Model (VGPM) of Behrenfeld and
cessed CZCS data used by Gregg et al. (2002) then led to a slight Falkowski (1997) was computed at 1 1 km2 spatial resolution
(10%) overestimate of global average concentrations. Fig. 6 shows from the MODIS Aqua data downloaded from the NASA GSFC.
the chlorophyll-a concentration climatologies derived from the MODIS daily surface chlorophyll-a concentrations, MODIS SST,
three ocean color sensors (CZCS, SeaWiFS, and MODIS). We find and MODIS cloud-corrected incident daily photosynthetically
that the newly reprocessed CZCS data (2006 NASA reprocessing) active radiation (PAR) at 1 1 km2 resolution were used as input
overestimates CHL within the subtropics (i.e. not just the Gulf of data. Monthly means for each parameter were used to calculate
Mexico) by a factor that can exceed 1.5 during winter. This overes- the weekly and monthly NPP. A monthly climatology (Fig. 7) and
timate of chlorophyll concentrations by CZCS is due to a consistent a series of monthly anomalies were then derived for each of the
bias toward lower CZCS remote sensing reflectance at 443 nm quadrants of the Gulf.
(Rrs443) and higher reflectance at 555 nm (Rrs555) relative to
SeaWiFS and MODIS. The differences are especially apparent dur- Climate indices
ing winter (NASA CZCS Technical Note, 2010).
We attempted to use the CZCS data by applying a bias correc- Two climate indices commonly used to help explain weather
tion based on a comparison of the CZCS and MODIS monthly mean patterns and oceanographic variability were examined to assess
chlorophyll-a concentration climatology, and subtracting the dif- whether changes in Gulf of Mexico observables were related to
ference from each CZCS pixel. However, the results (not shown) large-scale forcing. Specifically, we used the monthly Atlantic Mul-
rendered the CZCS data unusable for any quantitative analysis. tidecadal Oscillation (AMO) index of Enfield et al. (2001) and the
The correction effectively forced the CZCS data to look identical monthly El Niño/Southern Oscillation Multivariate ENSO Index
to the MODIS imagery on average. In addition to destroying any (ENSO MEI; Wolter and Timlin; 2011). The AMO index is an area-
possible trends in time, it changed the spatial patterns inherent weighted average of the SST estimates over the North Atlantic
in the Gulf as captured by CZCS imagery. We decided against using (0–70°N). It reflects coherent natural variability in the North Atlan-
such bias correction tools. tic Ocean (Poore et al., 2009). We used the unsmoothed version of
Because of our inability to correct the CZCS data, these were not the AMO (1948–2013; http://www.esrl.noaa.gov/psd/data/time-
useful for long-term trend analyses. The CZCS data would need to series/AMO/). The ENSO MEI tracks an ocean–atmosphere coupling
be reprocessed after recalibrating the blue bands using clear water that affects variability in climate and weather conditions around
radiance estimates in different hemispheres for each season. The the globe.
Fig. 6. Chlorophyll-a concentration monthly climatology [mg Chl m 3] from four different satellite sensors for the four ROI’s defined for the interior Gulf of Mexico (Fig. 1).
Thin solid line: CZCS data (1978–1986). Thick solid line: SeaWiFS (1998–2010). Dotted line: MODIS (2003–2010).
2 1
Fig. 7. Net Primary Production climatology (NPP; [mg C m d ]) derived from MODIS observations (2003–2010) for the four ROI’s defined for the interior Gulf of Mexico
(Fig. 1).
Results buoys in the interior of the Gulf of Mexico show a similar seasonal
cycle in all four regions, with consistent timing of maxima and min-
Wind speed ima of the average winds across the entire Gulf of Mexico (Fig. 2).
The agreement between the CCMP and buoy wind data show that
The long-term monthly average wind intensity derived from the the CCMP data may be used to examine long-term patterns,
CCM blended product and point measurements at the various NDBC including possible trends, in the wind over the Gulf of Mexico.
Different regions showed only slightly different amplitudes in the order of 0.2–0.3 m s 1 per decade (i.e. the null hypothesis that the
climatological average wind intensity. They showed similar long- slope is zero can be rejected; p 0.025). There has been no signif-
term average maxima in wind intensities of 7 m s 1 in the icant change in wind intensity in summer months (July through
November–February timeframe. All areas also showed a decrease September, the period of wind minima). However, over November
in wind intensity from about March–April, and minima (<4 to to January periods, when seasonal winds are highest, winds have
5 m s 1) in the July–September timeframe. The minima of this sum- strengthened (>0.36 m s 1 per decade; significant at p 0.025)
mertime period occur typically in the eastern Gulf of Mexico in all areas except the southeastern quadrant of the Gulf of Mexico.
(winds < 4 m s 1 ROI 1 and 3). The standard deviation in the mean In the southeastern quadrant, the increase shows weaker statistical
monthly wind intensity computed between 1987 and 2011 was significance. Overall, all regions show an increase in winds of over
similar during November–January compared to July–September >0.72 m s 1 (>1.4 knots) over the last 20 years in winter-time
periods (i.e. annual range in SD 0.5–1.2 m s 1 for all regions). winds. The increase in wind intensity in the Gulf of Mexico region
Fig. 8 shows the monthly wind anomaly in each of the four ROIs. is consistent with the observation that average annual wind inten-
The trend in the anomalies was overlaid on the anomaly curve. All sity has increased gradually over most of the world’s ocean area
areas show a significant increase in wind intensity with time of the since 1991 (Young et al., 2011a,b).
Northeast
Northwest
Southeast
Southwest
Fig. 8. Wind speed monthly anomaly (1987–2011) relative to the CCM wind climatology for the Gulf of Mexico. From top to bottom, panels show the Northeast, Northwest,
Southeast, and Southwest quadrants (Fig. 1). The solid line drawn through the series shows the least squares regression (Nov 1987-Dec 2011) line for the respective ROIs,
which has a significant slope of 0.2 m s 1 per decade for all areas (See Table 1 for statistics). The upper broken line is the least squares regression of wind anomalies for
November–January periods, plotted one standard deviation above the mean of the wind intensity anomalies for this period. The lower broken line is the least squares
regression for July through September periods, plotted one standard deviation below the mean of the summertime wind intensity anomalies.
Sea surface temperature (SST) global upper ocean temperature since the 1870s and especially
since the 1960s (Roemmich et al., 2012; Levitus et al.,
The long-term observations show that the seasonal SST cycle 2001,2009), and since the mid 1980s in the Caribbean Sea and
throughout the interior of the Gulf of Mexico is coherent (Fig. 3). tropical Atlantic Ocean (Chollett et al., 2012; Scranton et al.,
All quadrants show minima in February–March and maxima 2014) and Gulf of Mexico (Muhling et al., 2012). In the Gulf of Mex-
around August. The seasonal SST amplitude change was largest ico, cooler winters relative to our seasonal climatology were more
in the western Gulf (22–23 °C in February–March to 29–30 °C frequent in the 1980s than in the 2000s. Warmer summers relative
in August). The SST minima in the southeastern Gulf of Mexico to the climatology were also more frequent in the 2000s compared
reached only 24–25 °C. The highest variability in SST is observed to the 1980s. Indeed, 2009–2012 showed the four warmest consec-
during January–March in the northern quadrants of the Gulf. utive summers of this record of the Gulf, with 2010 and 2011 also
The time series of monthly regional SST anomalies within each showing the coolest winters. On average, all quadrants show a sig-
of the quadrants of the Gulf of Mexico shows gradual warming nificant (p < 0.025) trend of between +0.17 and +0.3 °C per decade
over the period 1982–2012 (Fig. 9). This is consistent with ship- (Table 1). The anomalies during August–September periods (when
and buoy-based observations of a gradual increase in the average maxima in SST are typically observed) as well as those in
Northeast
Northwest
Southeast
Southwest
Fig. 9. Monthly SST anomaly (1981–2011) derived relative to Pathfinder SST v5.2 monthly climatology. The least squares regression (November 1981–December 2011) lines
are overlaid. All areas showed a slope of between 0.17 and 0.3 °C per decade (See Table 1 for statistics). The upper broken line is the least squares regression of SST
anomalies for August–September periods, plotted one standard deviation above the mean of the SST anomalies for this period. The lower broken line is the least squares
regression for February–March periods, plotted one standard deviation below the mean of the wintertime SST anomalies. The broken line overlaid on all graphs is the
trendline for the AMO, computed for the period 1981–2012, to overlap with the Pathfinder SST anomaly data.
February–March (months of SST minima) increased at similar rates Mixed layer depth (MLD)
(significant at p < 0.025). The rate of increase in winter tempera-
tures was offset somewhat by the anomalously cold winters of The mixed layer depth has an important influence on biological
2010 and 2011, which were especially severe in the northern quad- activity in the Gulf of Mexico. The surface mixed layer deepens in
rants. Even though 2010 featured the largest large-scale tempera- winter due to convective mixing and wind forcing, and becomes
ture amplitude variation between winter and summer (i.e. a range shallow in summer due to insolation and a weaker seasonal wind
>8 °C in the southern quadrants and >9 °C in northern quadrants) regime (Muller-Karger et al., 1991; Melo Gonzalez et al., 2000).
since the beginning of the 1980s, there was no significant trend The mixing affects the timing and intensity of phytoplankton
in the amplitude of the winter-summer SST difference over the blooming over large scales, which has important bottom-up eco-
observation period. Overall, the highest rates of warming were logical implications for fish and other organisms at higher trophic
observed in the northwestern Gulf of Mexico quadrant (i.e. ROI levels (see, for example, Platt et al., 2003; Platt et al., 2007;
2; Table 1). Between the 1980s and the decade starting with Muhling et al., 2011).
2010, this region gained nearly 1 °C in average annual SST. The ECCO2 simulations show a strong seasonal cycle in the MLD
that is of similar magnitude, variability, and phase throughout the
interior of the Gulf of Mexico (Fig. 5). All quadrants show MLD
maxima of about 90–120 m in February, and minima of about
Sea surface height anomaly (SSHA) 20 m from about May through October. The southeastern quadrant
shows a slightly higher average summer MLD (30 m). While we
Satellite altimetry observations of variations in ocean surface have no direct evidence of the causes for this difference, it is pos-
topography provide a measure of heat content, of mesoscale eddy sible that the deeper summertime MLD in this region are due to the
variability, and of the fundamental processes controlling geo- very warm surface water entering the Gulf from the Caribbean Sea.
strophic ocean currents. Specifically, oceanic pressure centers drive The upper water column in this sector may be further homoge-
ocean currents much like atmospheric pressure centers drive nized by turbulence caused by the interaction between the Yuca-
atmospheric winds. SSHA data also help track changes in ocean tan/Loop Current and the continental shelf areas of Yucatan
heat content on seasonal and longer timescales. In the Gulf of Channel. This affects the average MLD computed over the south-
Mexico, the time series of SSHA observations collected since eastern region in the 1/4° spatial resolution model.
1992 illustrates and helps quantify the Loop Current ring shedding The monthly average MLD estimates obtained with the ECCO-2
process when the Loop Current intrudes northward. model for the Gulf of Mexico were similar to the long-term
The seasonal SSHA cycle in the interior of the Gulf of Mexico is monthly average MLD estimates derived by Muller-Karger et al.
spatially coherent throughout the region (Fig. 4). All quadrants (1991) using all NOAA NODC station data collected in Gulf of Mex-
show SSHA minima by late March and SSHA maxima by late ico waters deeper than 50 m between 1914 and 1985. They are also
August. The seasonal SSHA amplitude was largest in the northern similar to those provided by a climatological simulation of the
Gulf ( 0.05 m in March to 0.1 m in August, i.e. an amplitude region by Ezer (2000). Ezer found that his mixed layer depth sim-
of 0.15 m) and smallest in the southeastern Gulf ( 0.025 m in ulations were underestimated relative to field observations; he
March to 0.025 m in August, i.e. an amplitude of 0.05 m). The sig- also concluded that his model was rather insensitive to rapid
nal reflects steric changes due to seasonal warming of the water changes in wind speed (<6 h anomalies of opposite sign). In
column, but also the observation that over the course of the satel- general, the winter MLD computations based on field data of
lite altimetry time series, the highest frequency of Loop Current Muller-Karger et al. (1991; maximum winter MLD of the order of
anticyclonic ring-separation events occurred in August and Sep- 125–130 m) were slightly larger than those given by ECCO2
tember, with 7 and 4 events, respectively (Lindo-Atichati et al., (90–120 m). This difference can simply be attributed to a differ-
2013). ence in the criterion used to estimate the MLD, among many other
Fig. 10 shows the monthly SSHA anomaly in each of the four factors, and thus cannot be viewed as significant.
ROIs. The trend in the anomalies was overlaid on the anomaly There was no evidence of a systematic change in the MLD over
curve. The time series of monthly regional SSHA data within each time in any of the quarters of the Gulf (Fig. 11). The MLD anomaly
of the quadrants of the Gulf of Mexico shows a gradual increase showed no noticeable interannual variability during summer
in sea level variability over the period 1993–2012. All quadrants months, when the MLD is shallowest. There are some departures
show a significant increase of the order of 0.02–0.03 m per decade. from the mean during winter, usually in the range of 20–50 m
SSHA in the northeastern quadrant is more variable than in the and short-lived (i.e. one season or shorter). At the beginning of
southeastern, northwestern, and southwestern quadrants. This is 2010, a strong anomaly was observed during which MLD exceeded
due to the excursions of the LC into and out of the northeastern 50 m throughout the Gulf of Mexico for several weeks, coinciding
region. Also, starting in 2003, the LC was on average located more with the very cold winter that year.
to the north over the period of our records, and the average num-
ber of LC rings formed per year increased (Lindo-Atichati et al., Chlorophyll-a concentration and Net Primary Production
2013). The increase of SSHA in all quadrants follows the increase
of anticyclonic mesoscale activity in the eastern Gulf. The increase The chlorophyll-a concentration climatologies derived from
in SSHA in the western quadrants is lower than in the eastern each of the three ocean color sensors (CZCS, SeaWiFS, and MODIS)
quadrants, likely because only part of the warm waters entering show a clear seasonal pattern that is coherent throughout the deep
the Gulf from the Caribbean are transported to the west. The west- Gulf of Mexico (Fig. 6). The values and seasonal patterns found in
ward transport of SSHA occurs as the large anticyclonic Loop Cur- the reprocessed (2011) CZCS data are basically identical to those
rent rings migrate in that direction due to a b-plane effect (Shi and found by Muller-Karger et al. (1991) using the earlier calibration,
Nof, 1994; Nof, 2005). Given the strong relationship that exists atmospheric correction, and bio-optical algorithms from Gordon
between SSHA and the thermal structure of the ocean, our results et al. (1988). The CZCS measurements also show the same seasonal
suggest that the increase in mean SSHA in the interior of the GOM and regional patterns as those from SeaWiFS and MODIS, but the
over the last 20 years, which is especially marked since the early CZCS values are higher by a factor of 1.5 to 2. Because of this, the
2000s, is linked to the increase in the temperature of the upper CZCS data were not used to evaluate changes in phytoplankton
water column as shown in the SST observations. chlorophyll concentration between the 1980s and 2000s. The
66
Table 1
Trends in major satellite-observed parameters based on least-squares linear regression statistics.
Variable Region Climatological monthly Data years [inclusive] Anomaly trend (overalll) [var = A + B ⁄ time] (A, B) (R, N) [10 y delta] Anomaly trend (March–June) Anomaly trend (November–
of mean range [min–max] [var = A + B ⁄ time] (A, B) (R, N) [10 y January) [var = A + B ⁄ time] (A, B,
Interest delta] R, N)
SST (C) ROI 1 [22.6–29.6] 1981–2012 (A = 52.10, B = 0.026) (R = 0.37, N = 371) [10 y delta = 0.26] (A = 76.08, B = 0.038) (R = 0.46, (A = 26.96, B = 0.013) (R = 0.22,
N = 124) [10 y delta = 0.38] N = 93) [10 y delta = 0.13]
ROI 2 [21.9–29.7] 1981–2012 (A = 61.30, B = 0.030) (R = 0.44, N = 371) [10 y delta = 0.31] (A = 79.6, B = 0.04) (R = 0.50, (A = 38.54, B = 0.019) (R = 0.29,
N = 124) [10 y delta = 0.40] N = 93) [10 y delta = 0.19]
N = 124) [10 y delta = 0.22] N = 93) [10 y delta = 0.05]
N = 124) [10 y delta = 0.34] N = 93) [10 y delta = 0.14]
F.E. Muller-Karger et al. / Progress in Oceanography 134 (2015) 54–76

Wind (m/s) ROI 1 [4.0–7.4] 1987–2011 (A = 48.83, B = 0.024) (R = 0.24, N = 294) [10 y delta = 0.2] (A = 40.37, B = 0.020) (R = 0.2, (A = 72.62, B = 0.036) (R = 0.36,
N = 96) [10 y delta = 0.2] N = 74) [10 y delta = 0.4]
N = 96) [10 y delta = 0.2] N = 74) [10 y delta = 0.3]
N = 96) [10 y delta = 0.25] N = 74) [10 y delta = 0.2]
N = 96) [10 y delta = 0.29] N = 74) [10 y delta = 0.3]
SSHA (m) ROI 1 [ 0.03 to 0.1] 1993–2012 (A = 0.03, B = 0.00025), (R = 0.23, N = 238), [10 y delta = 0.03] (A = 0.006, B = 0.00005), (R = 0.06, (A = 0.01, B = 0.00001),
N = 238), [10 y delta = 0.006] (R = 0.06, N = 238), [10 y
delta = 0.01]
ROI 2 [ 0.07 to 0.07] (A = 0.01, B = 0.00013), (R = 0.16, N = 238), [10 y delta = 0.015] (A = 0.02, B = 0.0002), (R = 0.23, (A = 0.03, B = 0.00025),
N = 238), [10 y delta = 0.024] (R = 0.28, N = 238), [10 y
delta = 0.03]
ROI 3 [ 0.03 to 0.05] (A = 0.03, B = 0.00013), (R = 0.2, N = 238), [10 y delta = 0.016] (A = 0.02, B = 0.0001), (R = 0.22, (A = 0.02, B = 0.0002), (R = 0.21,
N = 238), [10 y delta = 0.012] N = 238), [10 y delta = 0.024]
ROI 4 [ 0.04 to 0.06] (A = 0.02, B = 0.00025), (R = 0.41, N = 238), [10 y delta = 0.2] (A = 0.03, B = 0.0002), (R = 0.49, (A = 0.02, B = 0.00015),
N = 238), [10 y delta = 0.024] (R = 0.31, N = 238), [10 y
delta = 0.018]
Chl-a ROI 1 CZCS: [0.17–0.46] 1979–2012 1979–1986: (A = 5.44, B = 0.0, R = 0, N = 88) 2003–2012: (A = 1.27, B = 0.0,
SeaWiFS/MODIS: R = 0, N = 180) 2009–2013: (A = 58.65, B = 0.03, R = 0.4, N = 49)
[0.12–0.25]
ROI 2 CZCS: [0.15–0.47] 1979–1986: (A = 1.0, B = 0.0, R = 0, N = 87) 2003–2012: (A = 0.9, B = 0.0,
SeaWiFS/MODIS: R = 0, N = 181) 2009–2013: (A = 28.90, B = 0.01, R = 0.3, N = 49)
[0.09–0.25]
ROI 3 CZCS: [0.10–0.25] 1979–1986: (A = 6.6, B = 0.0, R = 0.1, N = 90) 2003–2012: (A = 2.0, B = 0.0,
SeaWiFS/MODIS: R = 0.3, N = 181) 2009–2013: (A = 9.9, B = 0.0, R = 0.3, N = 49)
[0.07–0.19]
ROI 4 CZCS: [0.12–0.32] 1979–1986: (A = 12.2, B = 0.0, R = 0.2, N = 90) 2003–2012: (A = 1.2,
SeaWiFS/MODIS: B = 0.0, R = 0.1, N = 181) 2009–2013: (A = 14.54, B = 0.0, R = 0.3, N = 49)
[0.09–0.22]
MODIS NPP ROI 1 2003–2012 (A = 5.08, B = 0.09, R = 0.57, N = 121)
vs. ECCO2 ROI 2 2003–2012 (A = 9.37, B = 0.14, R = 0.81, N = 121)
MLD ROI 3 2003–2012 (A = 19.76, B = 0.28, R = 0.87, N = 121)
ROI 4 2003–2012 (A = 22.37, B = 0.22, R = 0.87, N = 121)
Northeast
Northwest
Southeast
Southwest
Fig. 10. Monthly SSHA anomaly (1993–2012), derived relative to AVISO SSHA monthly climatology. The least squares regression (January 1993–December 2012) lines are
overlaid. All areas showed a slope of between 0.015 and 0.03 m per decade (see Table 1 for statistics). The upper broken line is the least squares regression of SSH anomalies
for August–September periods, plotted one standard deviation above the mean of the SSHA anomalies for this period. The lower broken line is the least squares regression for
February–March periods, plotted one standard deviation below the mean of the wintertime SSHA anomalies.
SeaWiFS and MODIS time series, however, show no significant dec- and Salmerón-García et al., 2011). This peak was particularly con-
adal-scale trend in chlorophyll concentration over time in any of spicuous in the CZCS data, but it is still not clear whether this is
the regions of the Gulf (Fig. 12). simply an issue with the CZCS calibration and distortion of the esti-
Overall, SeaWiFS and MODIS show comparable concentration mated pigment concentrations.
values throughout the year in all regions, but SeaWiFS tends to In the northeastern quadrant, the secondary peak
overestimate the MODIS data by a few percent during the time (0.25 mg m 3) observed in July in the SeaWiFS and MODIS data,
of the year when concentrations are very low (<0.15 mg m 3), as and in August in the CZCS data (also higher values than SeaWiFS
during April through October. Highest concentrations (0.2– and MODIS), represents the offshore (east- and southeastward)
0.25 mg m 3) occurred on average in December and January, and dispersal of Mississippi River and other coastal waters from Missis-
minima (<0.1–0.15 mg m 3) in June through September. In general, sippi and Alabama (Muller-Karger et al., 1991; Gilbes et al., 1996;
the CZCS chlorophyll-a data had a significant positive bias (a factor Morey et al., 2005). We expected to see more variability in this
of 1.5–2) relative to SeaWiFS and MODIS. phenomenon, but instead it showed a strong seasonality, occurring
The two northern Gulf quadrants showed higher concentrations in 1982 and 1984 (CZCS) and in 1998, 1999, 2000, 2001, 2003,
and higher variability compared to the southern quadrants 2004, 2007, 2008, 2009, 2010, and 2011 (SeaWiFS and MODIS).
throughout the year. Chlorophyll-a concentrations in both the During the CZCS years, this event was most pronounced in August
northwestern and northeastern quadrants show secondary peaks (1979–1985). In the SeaWiFS and MODIS record, this was strongest
in the monthly climatology. In the northwestern Gulf of Mexico, in July. The timing of this phenomenon is associated with changes
a small regional peak occurs in May as a result of offshore transport in wind direction and intensity that cause offshore transport to the
of shelf waters. This is caused by the seasonal convergence of east and south (D’Sa and Korobkin, 2009; Salisbury et al., 2004;
waters coming from the Louisiana-Texas shelf in a counterclock- Muller-Karger, 2000). Once off the shelf, the plume can become
wise direction and from Mexico in a clockwise direction (Zavala- entrained in the Loop Current or its eddies (Muller-Karger et al.,
Hidalgo et al. 2003; Martínez-López and Zavala-Hidalgo, 2009; 1991; Del Castillo et al., 2000, 2001; Hu et al., 2005; Martínez-
Northeast
Northwest
Southeast
Southwest
Fig. 11. ECCO2 simulated mixed layer depth anomaly series (m). The least squares regression (January 1992–December 2012) lines are overlaid.
López and Zavala-Hidalgo, 2009). These offshore transports of The Net Primary Production (NPP) climatology for all four
coastal and shelf waters were especially marked in 1998, 2000, regions (Fig. 7) shows a peak in average productivity in February,
2008, 2009, and 2010. i.e. one to two months after the highest chlorophyll values are
To understand how the phytoplankton carbon to Chl-a ratio observed in all quadrants of the Gulf. Values of 700–
(C:Chl-a) may have changed with season, we explored possible 800 mg C m 2 d 1 are observed during this time in the northern
ranges in C:Chl-a ratios using the model of Cloern et al. (1995). quadrants of the Gulf. In the southwestern quadrant, the NPP peak
Since there are no published time series of either C:Chl-a or of is of the order of 600 mg C m 2 d 1, while it does not exceed
growth rates for the offshore Gulf of Mexico, we assumed possible 450 mg C m 2 d 1 in the southeastern quadrant. In all quadrants,
growth rates in the range of (l =) 0.2 and 1.0 d 1, SST extremes of the NPP peak occurs during the period of strongest winds (Novem-
22 and 30 °C, and nominal photosynthetically-active-radiation ber to February; Fig. 2), and SST minima (Fig. 3), which is when the
(PAR; i.e. ‘‘irradiance’’ in the Cloern model) of 1–12 mol maximum MLD occurs (Fig. 5). NPP minima in the northern half of
quanta m 2 d 1 in applying the model. Results show that there is the Gulf (300–350 mg C m 2 d 1) are reached in May. The NPP
likely a seasonality in C:Chl-a that coincides with that of Chl-a. minimum occurred when the MLD reached its lowest values. Low
Estimated C:Chl-a of 3 to 30 mg C (mg Chl-a) 1 were estimated NPP was maintained while the MLD was shallow, through October
for boreal summer (high temperatures and high irradiance, under (Fig. 5). The southwestern Gulf showed lower values
high or low growth rates). Higher C:Chl-a of about 5 to 42 mg C (250 mg C m 2 d 1) throughout this period. The lowest average
(mg Chl-a) 1 were estimated for winter (low temperatures and values (200 mg C m 2 d 1) occurred in the southeastern quad-
low irradiance, under high or low growth rates). rant. The NPP minimum period in general coincided with minima
Northeast
Northwest
Southeast
Southwest
Fig. 12. Chlorophyll-a concentration anomaly [mg m 3] series derived from SeaWiFS (1998–2002) and MODIS (2003–2013) data. The least squares regression (January 1998–
January 2013) lines for the respective ROIs are overlaid. None of the slopes were significantly different from zero.
in chlorophyll-a concentration in all areas. In the northeastern Peninsula (north of 24.5°N). For example, the positive anomaly of
quadrant, a substantial mid-summer secondary peak in NPP of January–March 2010 is the result of convective mixing due to cold
500 mg C m 2 d 1 was observed simultaneous with the July chlo- SST (SST anomalies of 1 °C) and strong winds observed Gulf-wide
rophyll-a peak. This may be an artifact in the NPP values related to at the end of 2009 and through January 2010 (see below). Some of
the regular offshore transport of coastal and riverine waters the positive anomalies in the southeastern region of the Gulf are
observed in July and August over the CZCS, SeaWiFS, and MODIS due to the high chlorophyll concentrations associated with the
time series, but, regardless, this is a highly productive region of upwelling phenomenon that begins along the western edge of
the Gulf in terms of fisheries. the Yucatan Current and propagates offshore along the periphery
As with chlorophyll concentration, there is no evidence of any of the Loop Current (Perez et al.; 1999a,b). This upwelling extends
long-term change in NPP values with time in any of the quadrants along the cyclonic edge of the Loop Current beyond the Yucatan
(Fig. 13). Since our regions of interest trace the 1000 m isobaths, shelf break. The bloom traces a sinuous arc that can extend over
variability in chlorophyll-a concentration in the southeastern and 1800 km into the northwestern and northeastern Gulf. SeaWiFS
southwestern quadrants reflects only blooms caused by deep and MODIS data show that this long bloom occurs most often dur-
ocean mixing or extreme cross-shelf transport events. In the south- ing January–March periods. This pattern was also clearly visible in
eastern quadrant, the time series of chlorophyll concentration the CZCS data. This bloom is advected and then likely sustained by
shows only small anomalies relative to the monthly climatology upwelling along the cyclonic edge of the Loop Current. Many satel-
(Fig. 12). This region experiences events north of the Yucatan lite images showed this bloom merging with the colored plume of
Northeast
Northwest
Southeast
Southwest
Fig. 13. Net Primary Production anomaly [mg C m 2 d 1] series derived from MODIS observations (2003–2013; black line) data. The least squares regression (January 2002–
January 2013) lines for the respective ROIs are overlaid. None of the slopes were significantly different from zero.
water advected offshore from the Mississippi River delta region as (between 1 and 1.5 °C relative to the climatology; Fig. 9). This
this plume became entrained along the northern edge of the Loop period experienced lower than average winds throughout the Gulf
Current, typically in the July timeframe. (0 to < 1 m s 1; Fig. 8). The conditions in 2011 likely led to shal-
At the end of 2009 and beginning of 2010, the entire Gulf of lower winter mixing relative to 2009–2010. The winter bloom of
Mexico experienced the largest positive chlorophyll-a anomalies 2010–2011 was also less pronounced relative to the one seen in
seen since the SeaWiFS record started in 1998 (Fig. 12). This large 2009–2010. These observations are consistent with the observa-
positive anomaly was synchronous throughout the Gulf. A much tion that increased mixing of the upper water column due to
smaller basin-wide positive anomaly was also observed in 1998– storms leads to higher chlorophyll-a concentrations in the deep
1999. The 2010 anomaly started as early as November–December Gulf of Mexico (Melo Gonzalez et al., 2000).
of 2009 and was most pronounced between January and March The 10-year MODIS NPP time series showed marked differences
2010. During the intervening months, the Gulf of Mexico experi- between the quadrants of the Gulf of Mexico, but it also showed
enced the coolest SST anomalies observed in the near three-dec- some periods during which changes occurred simultaneously
ade-long SST record (i.e. 1 to 2 °C relative to the climatology throughout the Gulf. All regions show a marked peak in the Janu-
for these months; Fig. 9). This coincided with stronger than average ary–March 2010 interval, with average NPP exceeding
basin-wide winds (sustained 0 to + 1 m s 1; Fig. 8). Another basin- 1400 mg C m 2 d 1 over the entire northern Gulf of Mexico in
wide cool anomaly, although somewhat less pronounced than the February that year (Fig. 13). Coherent minima (an anomaly of
one of 2010, was observed during January–February of 2011 200 mg C m 2 d 1) occurred in all quadrants of the Gulf in
February–March of 2003, March 2006, and January–February 2012. and NPP is in contrast to trends in physical variables that are
This last minimum was the most pronounced of the series. The fre- important to biological processes. Particularly, over the past
quent anomalies in January–February are driven in part by the 20 years we have seen an increase in sea surface temperature of
effect of the very large Gulf-wide peak observed in 2010 on the +0.6 °C, in the sea surface height anomaly of +0.06 m, and in win-
monthly climatologies. The pronounced NPP minimum of Janu- ter-time wind speed of up to +0.7 m s 1. The SeaWiFS observations
ary–February 2012 occurred when SST in the Gulf of Mexico were are calibrated, stable, and sufficiently accurate to conduct long-
high relative to the norm (+1 °C anomalies) and winds in the Gulf term, climate record analyses (Hu et al., 2001; Franz et al., 2007;
of Mexico were very weak (<6.5 m s 1 according to NOAA NDBC Eplee et al., 2011; Siegel et al., 2013). The MODIS calibration was
Buoy 42001 located in the Central Gulf; not shown). In contrast, cross-referenced to that of SeaWiFS. The chlorophyll-a and NPP
average NPP values were observed in each region for January– products used in this analysis were generated by the latest NASA
February 2011, a period that experienced equally weak winds reprocessing of these ocean color data (MODIS 2013 reprocessing).
but instead large SST negative anomalies (i.e. colder than 1 °C; Therefore, these results imply that, on average, annual phytoplank-
Figs. 8 and 9). ton concentration and productivity are insensitive to the long-term
changes in physical variables such as wind speed and SST observed
Climate indices over this period in the Gulf of Mexico. Below we explore a possible
reason for the lack of change observed over the past 15 years in
We found no significant correlation at the monthly scale average phytoplankton pigment concentration and Net Primary
between any of the parameters we examined in the Gulf of Mexico Production in the offshore Gulf of Mexico.
and either the monthly MEI or the monthly AMO indices, including The time series of chlorophyll values show very regular sea-
at lags spanning 0 to ±8 months. However, the AMO shows a posi- sonal changes in the southern quarters of the Gulf. Years with
tive long-term trend over the period of the satellite record exam- extremely cold winters, like the winter of 2009–2010, showed a
ined (1981–2012). While all quarters of the Gulf of Mexico positive anomaly in all areas. The cold temperatures observed in
showed positive trends in SST anomalies similar to that of the 2011 only seem to have led to markedly higher chlorophyll con-
AMO (Fig. 9), only the southeastern quadrant showed a trend that centrations only in the northern quadrants. We observed winds
was not significantly different to that of the AMO at the 95% level. during the 2009–2010 winter that were anomalously high in all
While there is no direct correlation between the monthly MEI and quadrants. Winds during winter of 2010–2011 were much milder,
the SST anomalies in the Gulf, the trend in the MEI over our study and showed a smaller to nil anomaly. In general, we observed
period between 1981 and 2012 was strongly negative (not shown). higher chlorophyll concentrations and primary productivity only
during periods when cooler temperature and higher winds coin-
Discussion cided. We saw normal levels of chlorophyll when winds were
average. And we measured low chlorophyll concentration when
Long-term changes in seasonal meteorological patterns, in the temperatures were warm and winds were weak at the same
average temperature of surface waters, or in the standing stock time.
and productivity of phytoplankton have important impacts on Melo Gonzalez et al., 2000 had already found positive anomalies
the habitat for numerous species, including humans. Long-term in pigment concentration in the Gulf of Mexico during winter
change can affect reproductive success, feeding habits, species months. This was associated with intense mixing of the water col-
interactions, migration routes, and ultimately define whether an umn by higher frequency and stronger winds associated with cold
organism may be present in a region at all or not. Studies such as fronts. They found a 4–5 month lag between low SST, high chloro-
that of Gable (1993) provide the earliest chart of the major ecolog- phyll concentrations, and the ENSO warm phase of 1982–1983
ical domains in the Gulf of Mexico, but these charts need to be (Enfield and Mayer, 1997) for the Caribbean and Gulf of Mexico.
redrafted frequently to help understand where change is taking Another positive pigment anomaly was observed in 1980–1981.
place. Long-term observations and baselines help develop the This anomaly was not related to ENSO, but rather to a period of
dynamic biogeographical regions needed to assess effects of distur- higher hurricane and extra-tropical low-pressure activity. Using a
bance due to natural events or anthropogenic causes including longer time series, we could not find a generalized relationship
resource development (Muller-Karger et al., 2014). between any of our parameters and the monthly ENSO MEI or
The waters of the interior of the Gulf of Mexico seaward of the AMO indices in the Gulf of Mexico at lags spanning 0 to +8 months.
continental margin continue to be seriously undersampled. A The positive trend in the AMO during the period overlapping the
search for observations of chlorophyll concentration in the NOAA Pathfinder SST data (1981–2012) is consistent with the AMO warm
NODC archives shows fewer than 100 chlorophyll observations phase that started in the mid-90s (Enfield and Cid-Serrano, 2010).
for each of the Regions of Interest defined for this study (Fig. 1). Over such time scales, the Gulf of Mexico reflects changes in heat
Most of these samples are associated with a particular cruise or balance taking place over the North Atlantic. The link between
with a seasonal program, such as the NOAA SEAMAP (SouthEast the North Atlantic and the Gulf of Mexico occurs through varia-
Area Monitoring and Assessment Program). We were not able to tions in the advective throughflow between the Yucatan Channel
derive a good chlorophyll concentration dataset from historical and the Straits of Florida (Liu et al., 2012). Because that flow is part
field observations archived at the NOAA NODC to compare with of the Atlantic Ocean meridional overturning circulation, the AMO
either CZCS, SeaWiFS, or MODIS chlorophyll estimates. Most sam- acts over timescales longer than a year. The negative trend in the
ples in the Gulf of Mexico available at the NODC are from the MEI over our study period is due to the repeated strong warm El
northern and eastern shelf regions, with relatively few samples Niño-Southern Oscillation (ENSO) events observed in the 1980s
available from offshore waters. Thus, information derived from and 1990s, and higher frequency of cooler ENSO events since
remote sensing is essential for characterization of the deep water 1999–2000, including the strong negative events of 1999–2000,
areas of the Gulf. 2007–2008, 2010–2011, and 2011–2012. Such extraneous factors
There was no significant trend in the chlorophyll-a concentra- (e.g., ENSO extremes) affect the annual heat budget of the Gulf of
tion time series for any of the regions in the Gulf of Mexico, either Mexico going into summer (Melo Gonzalez et al., 2000). They cause
during the CZCS years (1978–1986), or during the SeaWiFS and sporadic responses that have little or nothing to do with the North
MODIS years (1998–2012 inclusive). We also found no apparent Atlantic. Air-sea interactions driving changes in the Gulf of Mexico
trend in the NPP record. The lack of long term trends in chlorophyll at any particular time differ from those acting on average over the
North Atlantic. Whether the changes inferred from these two cli- Zavala-Hidalgo (2009) describe the temporal variability in appar-
mate indices are related is not yet clear. ent chlorophyll-a concentration in coastal and shelf waters around
In any event, the results from Muller-Karger et al. (1991) and the Gulf of Mexico and the mechanisms by which coastal waters
Melo Gonzalez et al., 2000 pointed to the depth of the mixed layer are advected across the shelf. The chlorophyll-a concentration cli-
as a controlling factor in defining chlorophyll-a levels in the Gulf of matology (Fig. 6) and the time series of anomalies (Fig. 12) clearly
Mexico. In this study we found the relationship between the MLD show that the largest deviations from the mean and the highest
estimates derived from the ECCO2 model and Chl-a concentration, temporal variability is observed in the northeastern quadrant of
and MLD and NPP, to be strong in every quadrant of the Gulf of the Gulf of Mexico. Muller-Karger et al. (1991), Muller-Karger
Mexico. Some outliers in this relationship stood out only in the (2000), Del Castillo et al. (2001), and Hu et al. (2005) had observed
northeastern quadrant. These results were consistent with the the offshore entrainment of Mississippi plume water into the inte-
Sverdrup critical depth hypothesis (Sverdrup, 1953). There is suffi- rior of the Gulf and subsequent advection to the southeast in the
cient light on a year-round basis to illuminate the mixed layer in Loop Current. We observed this entrainment in 11 out of the 14-
the Gulf of Mexico. Thus the ‘spring bloom’ occurs in the middle years of the series of SeaWiFS and MODIS images (1998–2012
of winter, when the mixed layer is deepest, allowing for a greater inclusive). It is this feature that leads to a different climatological
influx of nutrients into the euphotic zone, as Muller-Karger et al. chlorophyll pattern than expected for clear offshore waters, and
(1991) had already described. the cause for which Salmerón-García et al. (2011) and Callejas-
The statistical relationship between NPP and MLD is slightly Jimenez et al. (2012) identified the northeastern Gulf of Mexico
stronger than that of chlorophyll-a concentration. Table 1 shows as a uniquely different deep-water region in the Gulf.
the coefficients derived from a linear regression analysis per- Cross-shelf advection of coastal waters is forced by the seasonal
formed between NPP and MLD (Fig. 14). We chose to cast NPP as convergence of alongshore winds particularly to the southwest of
the independent variable and MLD as the dependent variable the Mississippi Delta, off the US-Mexico border region, and in the
because ocean color NPP observations are readily available, while Bay of Campeche (D’Sa and Korobkin, 2009; Martínez-López and
MLD estimates are difficult to obtain, particularly in a synoptic Zavala-Hidalgo; 2009). In the northeastern Gulf of Mexico, offshore
manner. We found that except in the northeast, NPP can predict advection of waters from the Mississippi River and other coastal
MLD well, explaining upwards of 70% of the variability in the waters is related to offshore transport due primarily to local wind.
MLD derived in the ECCO2 model. The relationship is less robust These waters can subsequently be entrained in the Loop Current or
in the northeastern quadrant, where high levels of chlorophyll-a its eddies (Muller-Karger, 2000), or in the southward drifts that
and NPP occurred regularly in July and August, during periods of occur over the west Florida shelf (Gilbes et al., 1996).
shallow MLD (Fig. 14). Visual examination of the satellite data time One question left unresolved by the observations reported here
series confirms that the high positive chlorophyll-a and NPP anom- is whether the effect of increasing stability of the upper water col-
alies observed in summer 1998, 2008, 2009, and 2011 in this umn due to rising surface temperatures is offset by the increase in
region were related to east and southward dispersal of Mississippi the anticyclonic mesoscale activity in the interior of the Gulf or by
water. the gradual increase in winds. The trend toward a warmer and
Muller-Karger et al. (1991), Biggs et al. (1993), Muller-Karger more windy Gulf of Mexico is not reflected in the biological indica-
(2000), Del Castillo et al. (2001), Salisbury et al. (2004), Hu et al. tors Chl-a and NPP. We do not yet have a conceptual model that
(2005), D’Sa and Korobkin (2009), and Martínez-López and can provide an explanation for which process may dominate. This
Northwest Northeast
Southwest Southeast
Fig. 14. Monthly mean MODIS-derived Net Primary Production (NPP, [mg C m 2 day 1]) against monthly mean ECCO2 mixed layer depth (MLD, [m]) in each of the Gulf of
Mexico quarters. Parameters for the least-squares linear regression are shown in Table 1.
question is relevant for assessing impacts of climate change over References

both regional and global scales.
The increase in oceanic sea surface temperature has already Adams, C.M., Hernandez, E., Cato, J.C., 2004. The economic significance of the Gulf of
Mexico related to population, income, employment, minerals, fisheries and
been highlighted as a driver of a possible increase in hurricane shipping. Ocean and Coastal Management 47 (11–12), 565–580. http://
activity (Goldenberg et al., 2001). More generally, the trend toward dx.doi.org/10.1016/j.ocecoaman.2004.12.002.
stronger winds over large areas of the ocean (Young et al., 2011a,b) Alvera-Azcárate, A., Barth, A., Weisberg, R.H., 2009. The surface circulation of the
Caribbean Sea and the Gulf of Mexico as inferred from satellite altimetry.
and higher upper ocean temperatures (Levitus et al., 2001; Journal of Physical Oceanography 39 (3), 640–657.
Roemmich et al., 2012) are linked through intensification of the Atlas, R., Hoffman, R.N., Ardizzone, J., Leidner, S.M., Jusem, J.C., Smith, D.K., Gombos,
Walker and Hadley cells in the atmosphere (England et al., 2014). D., 2011. A cross-calibrated, multiplatform ocean surface wind velocity product
for meteorological and oceanographic applications. Bulletin of the American
England et al. (2014) attribute cooling of surface waters in the
Meteorological Society 92, 157–174. http://dx.doi.org/10.1175/
equatorial Pacific to the intensification of the Trade Winds 2010BAMS2946.1.
observed in the Pacific in the 2001–2012 timeframe. Baringer, M.O.N., Larsen, J.C., 2001. Sixteen years of Florida current transport at
The significant warming and increase in the sea surface height 27°N. Geophysical Research Letters 28 (16), 3179–3182. http://dx.doi.org/
10.1029/2001GL013246.
anomaly that we observe in the interior of the Gulf also helps to Behrenfeld, M.J., Falkowski, P.G., 1997. Photosynthetic rates derived from satellite –
explain the rising coastal sea levels observed by Hanson and based chlorophyll concentration. Limnology and Oceanography 42, 1–20.
Maul (1993), Maul et al. (2001), and Wahl et al. (2014). Part of Behringer, D.W., Molinari, R.L., Festa, J.F., 1977. The variability of anticyclonic
current patterns in the Gulf of Mexico. Journal of Geophysical Research 82,
the increase in the sea surface height anomaly is a direct conse- 5469–5476.
quence of steric changes due to this warming of the water column. Biggs, D.C., Muller-Karger, F.E., 1993. Ship and satellite observations of chlorophyll
Sea level rise due to climate change is one of the most important stocks in warm- and cold-core rings in the western Gulf of Mexico. Journel of
Geophysical Research 99 (C4), 7371–7384.
threats to ecosystem services provided at present by coastal areas Biggs, D.C., Jochens, A.E., Howard, M.K., DiMarco, S.F., Mullin, K.D., Leben, R.R.,
of the Gulf of Mexico (Twilley et al., 2001). Yet, how this affects Muller-Karger, F.E., Hu, C., 2005. Eddy forced variations in on- and off-margin
basic processes like primary production in the interior of the Gulf summertime circulation along the 1000-m Isobath of the Northern Gulf of
Mexico, 2000–2003, and links with sperm whale distributions along the middle
of Mexico is not clear. We know that changes in mesoscale eddy shelf. In: Sturges, W., Lugo-Fernandez, A. (Eds.), New Developments in the
variability play an important role in providing nutrients for phyto- Circulation of the Gulf of Mexico. AGU Geophysical Monograph Series 161, pp.
plankton growth (McGillicuddy et al., 1999; Mahadevan et al., 71–86.
Bunge, L., Ochoa, J., Badan, A., Candela, J., Sheinbaum, J., 2002. Deep flows in the
2012, 2010). Piontkovski and Nezlin (2012) found a positive corre-
Yucatan Channel and their relation to changes in the loop current extension.
lation between negative SSHA and chlorophyll concentration in the Journal of Geophysical Research 107.
western Arabian Sea, emphasizing that chlorophyll was generally Callejas-Jimenez, M., Santamaria-del-Angel, E., Gonzalez-Silvera, A., Millan-Nuñez,
lower in years where SSHA was more positive (i.e. anticyclonic R., Cajal-Medrano, R., 2012. Dynamic regionalization of the Gulf of Mexico
based on normalized radiances (nLw) derived from MODIS-Aqua. Continental
eddies dominate) than negative (i.e. cyclonic eddies dominate). In Shelf Research 37, 8–14. http://dx.doi.org/10.1016/j.csr.2012.01.014.
the Canary Eddy Corridor, chlorophyll concentration may be Candela, J., Tanahara, S., Crepon, M., Barnier, B., Sheinbaum, J., et al., 2003. Yucatan
enhanced by both anticyclonic and cyclonic eddies (Sangrà et al., Channel flow: observations versus CLIPPER ATL6 and MERCATOR PAM models.
Journal of Geophysical Research 108 (C12), 24pp. http://dx.doi.org/10.1029/
2009). Examining the hypothesis that anticyclonic or cyclonic 2003JC001961.
eddies, or both, have an impact on chlorophyll requires a synoptic Cardona, Y., Bracco, A., 2014. Predictability of mesoscale circulation throughout the
eddy census in the Gulf of Mexico, which was outside the scope of water column in the Gulf of Mexico. Deep Sea Research Part II: Topical Studies
in Oceanography. http://dx.doi.org/10.1016/j.dsr2.2014.01.008.
the present study. Again, these questions are relevant to global Casey, K.S., Brandon, T.B., Cornillon, P., Evans, R., 2010. The past, present and future
change science because these processes apply to regional seas gen- of the AVHRR Pathfinder SST program. In: Barale, V., Gower, J.F.R., Alberotanza,
erally and the impacts of concurrent change are important at basin L. (Eds.), Oceanography From SPACE. Springer. http://dx.doi.org/10.1007/978-
90-481-8681-5_16.
and global spatial scales. Cerdeira-Estrada, S., Muller-Karger, F.E., Gallegos Garcia, A., 2005. Variability of the
The results that there has been no change over the past 20 years sea surface temperature around cuba. Gulf of Mexico Science 23 (2), 161–171.
in average mixed layer depth is thus very consistent with the Chang, Y.-L., Oey, L.-Y., 2010. Eddy and wind-forced heat transports in the Gulf of
Mexico. Journal of Physical Oceanography 40, 2728–2742. http://dx.doi.org/
observation that there have been no long-term trends in chloro-
10.1175/2010JPO4474.1.
phyll concentration or in the overall phytoplankton productivity Chang, Y.L., Oey, L.Y., 2012. Why does the Loop Current tend to shed more eddies in
of offshore waters of the Gulf. However, answering the question summer and winter? Geophysical Research Letters 39 (5).
of whether the increase in the wind stress can offset the decrease Chang, Y.L., Oey, L.Y., 2013a. Loop Current growth and eddy shedding using models
and observations: numerical process experiments and satellite altimetry data.
in regional primary productivity that would be expected solely Journal of Physical Oceanography 43 (3), 669–689.
from increased thermal stability of surface waters probably Chang, Y.L., Oey, L.Y., 2013b. Coupled response of the trade wind, SST Gradient, and
requires a detailed study using a dynamic mixed layer model. SST in the Caribbean Sea, and the potential impact on Loop Current’s
interannual variability. Journal of Physical Oceanography 43 (7), 1325–1344.
We encourage such a study as this is at the core of the future health Chérubin, L.M., Sturges, W., Chassignet, E.P., 2005. Deep flow variability in the
of the Gulf of Mexico ecosystem. vicinity of the Yucatan Straits from a high-resolution numerical simulation.
Journal of Geophysical Research – Part C-Oceans 110, C04009.
Chollett, I., Müller-Karger, F.E., Heron, S.F., Skirving, W., Mumby, P.J., 2012. Seasonal
Acknowledgements and spatial heterogeneity of recent sea surface temperature trends in the
Caribbean Sea and southeast Gulf of Mexico. Marine Pollution Bulletin. http://
dx.doi.org/10.1016/j.marpolbul.2012.02.016.
This work was supported by ExxonMobil Upstream Research Cloern, J.E., Grenz, C., Vidergar-Lucas, L., 1995. An empirical model of the
Company, NASA grant NNX14AP62A ‘National Marine Sanctuaries phytoplankton chlorophyll:carbon ratio-the conversion factor between
as Sentinel Sites for a Demonstration Marine Biodiversity Observa- productivity and growth rate. Limnology and Oceanography 40 (7), 1313–1321.
Conkright, M.E., Locarnini, R.A., Garcia, H.E., O’Brien, T.D., Boyer, T.P., Stephens, C.,
tion Network (MBON)’ funded under the National Ocean Partner- Antonov, J.I., 2002. World Ocean Atlas 2001: Objective Analyses, Data Statistics,
ship Program (NOPP RFP NOAA-NOS-IOOS-2014-2003803 in and Figures, CD-ROM Documentation. National Oceanographic Data Center,
partnership between NOAA, BOEM, and NASA), NASA grant Silver Spring, MD, 17 pp.
Del Castillo, C., Gilbes, F., Coble, P., Muller-Karger, F.E., 2000. On the dispersal of
NNX11AP76G ‘Management and conservation of Atlantic Bluefin riverine colored dissolved organic matter over the West Florida Shelf.
tuna (Thunnus thynnus) and other highly migratory fish in the Gulf Limnology and Oceanography 45 (6), 1425–1432.
of Mexico under IPCC climate change scenarios’, and NOAA (AOML/ Del Castillo, C., Coble, P., Conmy, R., Muller-Karger, F.E., Vanderbloemen, L., Vargo,
G., 2001. Multispectral in-situ measurements of organic matter and chlorophyll
CIMAS Grant ‘Ocean Conditions in The Gulf of Mexico’). Ocean cir- fluorescence in seawater: documenting the intrusion of the Mississippi River
culation simulation results were kindly provided by the Estimating Plume in the West Florida Shelf. Limnology and Oceanography 46 (7), 1836–
the Circulation and Climate of the Ocean, Phase II (ECCO2) Project. 1843.
Ding, K., Gordon, H.R., 1995. Analysis of the influence of O2 A-band absorption on observing system to detect sea surface temperature anomaly events in the
atmospheric correction of ocean-color imagery. Applied Optics 34, 2068–2080. Florida keys. IEEE Transactions on Geoscience and Remote Sensing 47 (6),
D’Sa, E.J., Korobkin, M., 2009. Examining SeaWiFS chlorophyll variability along the 1607–1620, June 2009.
Louisiana coast using wavelet analysis. In: Proceedings, OCEANS 2009, MTS/ Hurlburt, H.E., Thompson, J.D., 1980. A numerical study of loop current intrusions
IEEE Biloxi – Marine Technology for Our Future: Global and Local Challenges, 1– and eddy shedding. Journal of Physical Oceanography 10, 1611–1651.
5. Kalnay, E., Kanamitsu, M., Kistler, R., Collins, W., Deaven, D., Gandin, L., Iredell, M.,
Enfield, D.B., Cid-Serrano, L., 2006. Projecting the risk of future climate shifts. Saha, S., White, G., Woollen, J., Zhu, Y., Chelliah, M., Ebisuzaki, W., Higgins, W.,
International Journal of Climatology 26 (7), 885–895. Janowiak, J., Mo, K.C., Ropelewski, C., Wang, J., Leetmaa, A., Reynolds, R., Jenne,
Enfield, D.B., Cid-Serrano, L., 2010. Secular and multidecadal warmings in the North Roy, Joseph, Dennis, 1996. The NCEP/NCAR 40-year reanalysis project. Bulletin
Atlantic and their relationships with major hurricane activity. International of the American Meteorological Society 77 (3), 437–471.
Journal of Climatology 30, 174–184. Kara, A.B., Rochford, P.A., Hurlburt, H.E., 2000. An optimal definition for ocean
Enfield, D.B., Mayer, D.A., 1997. Tropical Atlantic SST variability and its relation to El mixed layer depth. Journal of Geophysical Research 105 (C7), 16803–16821.
Niño-Southern oscillation. Journal of Geophysical Research 102, 929–945. http://dx.doi.org/10.1029/2000JC900072.
Enfield, D.B., Mestas-Nunez, A.M., Trimble, P.J., 2001. The Atlantic multidecadal Karnauskas, M., Schirripa, M.J., Kelble, C.R., Cook, G.S., Craig, J.K., 2013. Ecosystem
oscillation and its relationship to rainfall and river flows in the continental U.S. status report for the Gulf of Mexico. NOAA Technical Memorandum NMFS-
Geophysical Research Letters 28, 2077–2080. SEFSC-653.
England, M.H., McGregor, S., Spence, P., Meehl, G.A., Timmermann, A., Cai, W., Sen Kennedy, A.J., Griffin, M.L., Morey, S.L., Smith, S.R., O’Brien, J.J., 2007. Effects of El
Gupta, A., McPhaden, M.J., Purich, A., Santoso, A., 2014. Recent intensification of Niño–southern oscillation on sea level anomalies along the Gulf of Mexico
wind-driven circulation in the Pacific and the ongoing warming hiatus. Nature coast. Journal of Geophysical Research: Oceans 112 (C5), 112. http://dx.doi.org/
Climate Change 4. http://dx.doi.org/10.1038/NCLIMATE2106. 10.1029/2006JC003904.
Eplee, R.E., Jr., Meister, G., Patt, F.S., Franz, B.A., McClain, C.R., 2011. Uncertainty Kistler, R., Collins, W., Saha, S., White, G., Woollen, J., Kalnay, E., Chelliah, M.,
assessment of the SeaWiFS on-orbit calibration. In: Butler, J.J., Xiong, X., Gu, X. Ebisuzaki, W., Kanamitsu, M., Kousky, V., van den Dool, H., Jenne, R., Fiorino, M.,
(Eds.), Earth Observing Systems XVI. Proc. SPIE, 8153. pp. 815310. 2001. The NCEP–NCAR 50–year reanalysis: monthly means CD–ROM and
Etter, P.C., 1983. Heat and freshwater budgets of the Gulf of Mexico. Journal of documentation. Bulletin of the American Meteorological Society 82 (2), 247–
Physical Oceanography 13, 2058–2069. 267 (February 2001).
Ezer, T., 2000. On the seasonal mixed-layer simulated by a basin-scale ocean model Laws, E.A., Bannister, T.T., 1980. Nutrient- and light-limited growth of Thalassiosira
and the Mellor-Yamada turbelence scheme. Journal of Geophysical Research fluviatilis in continuous culture, with implications for phytoplankton growth in
105 (C7), 16843–16855. the ocean. Limnol and Oceanography 25, 457–473.
Ezer, T., Oey, L.-Y., Lee, H.-C., Sturges, W., 2003. The variability of currents in the Levitus, S., Antonov, J.L., Wang, J., Delworth, T.L., Dixon, K.W., Broccoli, A.J., 2001.
Yucatan Channel: analysis of results from a numerical Ocean model. Journal of Anthropogenic warming of Earth’s climate system. Science 292,
Geophysical Research 108 (C1), 3012. http://dx.doi.org/10.1029/2002JC001509. 267–270.
Falkowski, P.G., Dubinsky, Z., Wyman, K., 1985. Growth-irradiance relationships in Levitus, S., Antonov, J.I., Boyer, T.P., Locarnini, R.A., Garcia, H.E., Mishonov, A.V.,
phytoplankton. Limnology and Oceanography 30, 311–321. 2009. Global ocean heat content 1955–2008 in light of recently revealed
Felder, D.L., Camp, D.K. (Eds.), 2012. Gulf of Mexico Origin, Waters, and Biota instrumentation problems. Geophysical Research Letters 36, L07608. http://
Biodiversity. Texas A&M University Press. dx.doi.org/10.1029/2008GL037155.
Franz, B.A., Bailey, S.W., Werdell, P.J., McClain, C.R., 2007. Sensor-independent Le Hénaff, M., Kourafalou, V.H., Morel, Y., Srinivasan, A., 2012. Simulating the
approach to vicarious calibration of satellite ocean color radiometry. Applied dynamics and intensification of cyclonic Loop Current Frontal Eddies in the Gulf
Optics 46, 5068–5082. of Mexico. Journal of Geophysical Research: Oceans (1978–2012) 117 (C2).
Fratantoni, P.S., Lee, T.N., Podesta, G.P., Muller-Karger, F., 1998. The influence of Le Traon, P.Y., Nadal, F., Ducet, N., 1998. An improved mapping method of multi-
Loop Current variability on the formation and evolution of cyclonic eddies in satellite altimeter data. Journal of Atmospheric and Oceanic Technology 15,
the Southern Straits of Florida. Journal of Geophysical Research 103 (C11), 522–534.
24759–24779. Li, Q.P., Franks, P.J.S., Landry, M.R., Goericke, R., Taylor, A.G., 2010. Modeling
Gable, F.J., 1993. Marine habitats: selected environmental and ecological charts. In: phytoplankton growth rates and chlorophyll to carbon ratios in California
Maul, George (Ed.), Climate Change in the Intra-Americas Sea. United Nations coastal and pelagic ecosystems. Journal of Geophysical Research 115, G04003.
Environmental Programme (Wider Caribbean Region) and Intergovernmental http://dx.doi.org/10.1029/2009JG001111.
Oceanographic Commissions (Caribbean and Adjacent Regions). Edward Arnold Lindo-Atichati, D., Bringas, F., Goni, G., 2013. Loop Current excursions and ring
Publishers, pp. 217–261 (Chapter 10). detachments during 1993–2009. International Journal of Remote Sensing 34
Gilbes, F., Tomas, C., Walsh, J.J., Muller-Karger, F.E., 1996. An episodic chlorophyll (14), 5042–5053. http://dx.doi.org/10.1080/01431161.2013.787504.
plume on the west Florida shelf. Continental Shelf Research 16 (9), 1201–1224. Lipinski, D., Mohseni, K., 2014. Observations on the flow structures and transport in
Goldenberg, S.B., Landsea, C.W., Mestas-Nunez, A.M., Gray, W.M., 2001. The recent a simulated warm-core ring in the Gulf of Mexico. Ocean Dynamics 64 (1), 79–
increase in Atlantic hurricane activity: causes and implications. Science 293, 88.
474–479. Liu, Y., Lee, S.-K., Muhling, B.A., Lamkin, J.T., Enfield, D.B., 2012. Significant reduction
Gordon, H.R., 1997. Atmospheric correction of ocean color imagery in the earth of the Loop Current in the 21st century and its impact on the Gulf of Mexico.
observing system era. Journal of Geophysical Research 102, 17081–17106. Journal of Geophysical Research, 117. http://dx.doi.org/10.1029/2011JC007555,
Gordon, H.R., Wang, M., 1994. Retrieval of water-leaving radiance and aerosol C05039.
optical thickness over the oceans with SeaWiFS: a preliminary algorithm. Luis, Capurro, R.A., Reid, Joseph L. (Eds.), 1972. Contributions on the Physical
Applied Optics 33, 443–452. Oceanography of the Gulf of Mexico. volume 2 of Texas A & M University
Gordon, H.R., Brown, O.B., Evans, R.H., Brown, J.W., Smith, R.C., Baker, K.S., Clark, oceanographic studies, Texas A & M University Dept. of Oceanography. The
D.K., 1988. A semianalytic radiance model of ocean color. Journal of Geophysical University of California Press.
Research 93 (D9), 10909–10924. Mahadevan, A., Tandon, A., Ferrari, R., 2010. Rapid changes in mixed layer
Gregg, W.W., Conkright, M.E., 2001. Global seasonal climatologies of ocean stratification driven by submesoscale instabilities and winds. Journal of
chlorophyll: blending in situ and satellite data for the CZCS era. Journal of Geophysical Research 115, C03017. http://dx.doi.org/10.1029/2008JC005203.
Geophysical Research 106, 2499–2515. Mahadevan, A., D’Asaro, E., Lee, C., Perry, M.-J., 2012. Eddy-driven stratification
Gregg, W.W., Conkright, Margarita E., O’Reilly, John E., Patt, Frederick S., Wang, initiates North Atlantic spring phytoplankton blooms, 2012. Science 337 (6090),
Menghua H., Yoder, James A., Casey, Nancy W., 2002. NOAA–NASA coastal zone 54–58. http://dx.doi.org/10.1126/science.1218740.
color scanner reanalysis effort. Applied Optics 41 (9), 1615–1628. Maltrud, M., Peacock, S., Visbeck, M., 2010. On the possible long-term fate of oil
Hanson, K., Maul, G.A., 1993. Analysis of temperature, precipitation, and sea level released in the Deepwater Horizon incident, estimated using ensembles of dye
variability with concentration on Key West, Florida, for evidence of trace-gas- release simulations. Environmental Research Letters 5, 035301. http://
induced climate change. In: Maul, George (Ed.), Climate Change in the Intra- dx.doi.org/10.1088/1748-9326/5/3/035301.
Americas Sea. United Nations Environmental Programme (Wider Caribbean Marshall, J., Adcroft, A., Hill, C., Perelman, L., Heisey, C., 1997a. A finite-volume,
Region) and Intergovernmental Oceanographic Commissions (Caribbean and incompressible Navier-Stokes model for studies of the ocean on parallel
Adjacent Regions). Edward Arnold Publishers, pp. 193–213 (Chapter 9). computers. Journal of Geophysical Research 102 (C3), 5753–5766.
Herring, J.H., 2010. Gulf of Mexico hydrographic climatology and method of Marshall, J., Hill, C., Perelman, L., Adcroft, A., 1997b. Hydrostatic, quasi-hydrostatic,
synthesizing subsurface profiles from the satellite sea surface height anomaly. and nonhydrostatic ocean modeling. Journal of Geophysical Research 102 (C3),
US. Department of Commerce. National Oceanic and Atmospheric 5733–5752.
Administration. National Ocean Service. Coastal Survey Development Martínez-López, B., Zavala-Hidalgo, J., 2009. Seasonal and interannual variability of
Laboratory. Report 122, 63 pp. cross-shelf transports of chlorophyll in the Gulf of Mexico. Journal of Marine
Hu, C., Carder, K.L., Muller-Karger, F.E., 2001. How precise are SeaWiFS ocean color Systems 77 (2009), 1–20. http://dx.doi.org/10.1016/j.jmarsys.2008.10.002.
estimates? Implications of digitization-noise errors. Remote Sensing of Maul, G.A., 1978. The 1972–1973 cycle of the Gulf Loop current. Part II: Mass and
Environment 76, 239–249. salt balances of the basin. Symposium on Progress in Marine Research in the
Hu, C., Nelson, J.R., Johns, E., Chen, Z., Weisberg, R.H., Muller-Karger, F.E., 2005. Caribbean and Adjacent Regions. Caracas (Venezuela), 12 July 1976. FAO
Mississippi River water in the Florida Straits and in the Gulf Stream off Georgia Fisheries Reports (FAO). No. 200 (suppl.), pp. 597–619.
in summer 2004. Geophysical Research Letters 32 (10), L14606. http:// Maul, G.A., Mayer, D.A., Baig, S.R., 1985. Comparison between a continuous three-
dx.doi.org/10.1029/2005GL022942. year current meter observation at the sill of the Yucatan Strait, satellite
Hu, C., Muller-Karger, F., Murch, B., Myhre, D., Taylor, J., Luerssen, R., Moses, C., measurements of Gulf Loop Current area, and regional sea level. Journal of
Zhang, C., Gramer, L., Hendee, J., 2009. Building an automated integrated Geophysical Research 90, 9089–9096.
Maul, G.A., Davis, A.M., Simmons, J.W., 2001. Seawater temperature trends at USA Postlaunch Calibration and Validation Analyses, Part 3. NASA Tech. Memo.
tide gauges. Geophysical Research Letters 28 (20), 3935–3937. 2000–206892, vol. 11. NASA Goddard Space Flight Center, Greenbelt, Maryland,
McGillicuddy Jr., D.J., Johnson, R., Siegel, D.A, Michaels, A.F., Bates, N.R., Knap, A.H., pp. 9–23.
1999. Mesoscale variations of biochemical properties in the Sargasso Sea. Ortner, P.B., Lee, T.N., Milne, P.J., Zika, R.G., Clarke, M.E., Podesta, G.P., Swart, P.K.,
Journel of Geophysical Research 104 (C6), 13381–13394. http://dx.doi.org/ Tester, P.A., Atkinson, L.P., Johnson, W.R., 1995. Mississippi River flood waters
10.1029/1999JC900021. that reached the Gulf Stream. Journal of Geophysical Research 100 (C7), 13595–
Melo Gonzalez, N., Muller-Karger, F.E., Cerdeira Estrada, S., Pérez de los Reyes, R., 13601. http://dx.doi.org/10.1029/95JC01039.
Victoria del Rio, I., Cárdenas Perez, P., Mitrani Arenal, I., 2000. Near-surface Perez, R., Muller-Karger, F.E., Merino, M., Victoria, I., Melo, N., Cerdeira, S., 1999a.
phytoplankton distribution in the western Intra-Americas Sea: The influence of Cuban, Mexican, US researchers probing mysteries of Yucatan current. Earth In
El Niño and weather events. Journal of Geophysical Research 105 (C6), 14029– Space 12 (1), 10–14, September 1999.
14043. Perez, R., Muller-Karger, F.E., Victoria, I., Melo, N., Cerdeira, S., 1999b. Cuban,
Menemenlis, D., Fukumori, I., Lee, T., 2005a. Using greens functions to calibrate an Mexican, US researchers probing mysteries of Yucatan current. EOS. AGU
ocean general circulation model. Monthly Weather Review 133, 1224–1240. Transactions 80 (14), 153.
Menemenlis, D., Hill, C., Adcroft, A., Campin, J., Cheng, B., Ciotti, B., Fukumori, I., Piontkovski, S., Nezlin, N., 2012. Mesoscale eddies of arabian Sea: physical–
Heimbach, P., Henze, C., Koehl, A., Lee, T., Stammer, D., Taft, J., Zhang, J., 2005b. biological interactions. International Journal of Marine Science 2 (7), 51–56.
NASA supercomputer improves prospects for ocean climate research. Eos, http://dx.doi.org/10.5376/ijms. 2012.02.0007.
Transactions American Geophysical Union 86 (89), 95–96. Platt., T., Fuentes-Yaco, C., Frank, K.T., 2003. Spring algal bloom and larval fish
Menemenlis, D., Campin, J., Heimbach, P., Hill, C., Lee, T., Nguyen, A., Schodlock, M., survival. Nature 423.
Zhang, H., 2008. ECCO2: high resolution global ocean and sea ice data synthesis. Platt, T., Sathyendranath, S., Fuentes-Yaco, C., 2007. Biological oceanography and
Mercator Ocean Quarterly Newsletter 31, 13–21. fisheries management: perspective after 10 years. Ices Journal of Marine
Menzel, D.W., Ryther, J.H., 1961. Annual variations in primary production of the Science. 64 (5), 863–869.
Sargasso Sea off Bermuda. Deep-Sea Research 7, 282–288. Poore, R.Z., DeLong, K., Richey, J., Quinn, T.M., 2009. Evidence of multidecadal
Mildner, T.C., Eden, C., Czeschel, L., 2013. Revisiting the relationship between Loop climate variability and the Atlantic multidecadal oscillation from a Gulf of
Current rings and Florida current transport variability. Journal of Geophysical Mexico sea-surface temperature-proxy record. Geo-Marine Letters 29 (6), 477–
Research: Oceans 118, 1–10. http://dx.doi.org/10.1002/2013JC009109. 484.
Morel, A., Prieur, L., 1977. Analysis of variations in ocean color. Limnology and Rio, M.-H., Hernandez, F., 2004. A mean dynamic topography computed over the
Oceanography 22, 709–722. world ocean from altimetry, in situ measurements, and a geoid model. Journal
Morey, S.L., Zavala-Hidalgo, J., O’Brien, J.J., 2005. The Seasonal variability of of Geophysical Research 109, C12032. http://dx.doi.org/10.1029/2003JC002226.
continental shelf circulation in the northern and Western Gulf of Mexico from Rivas, D., Badan, A., Ochoa, J., 2005. The ventilation of the deep Gulf of Mexico.
a high-resolution numerical model. In: Circulation in the Gulf of Mexico: Journal of physical oceanography 35, 1763–1781.
Observations and Models. Geophysical Monograph Series 161. American Roemmich, D., Gould, W.J., Gilson, J., 2012. 135 years of global ocean warming
Geophysical Union, pp. 203–218. http://dx.doi.org/10.1029/161GM16. between the Challenger expedition and the Argo Programme. Nature Climate
Morrison, Nowlin, W.D., 1977. Repeated nutrient, oxygen, and density sections Change. http://dx.doi.org/10.1038/nclimate1461.
through the Loop Current. Journal of Marine Research 35 (1), 105–128. Ryther, J.H., Menzel, D.W., 1960. The seasonal and geographical range of primary
Muhling, B.A., Lee, S.-K., Lamkin, J.T., Liu, Y., 2011. Predicting the effects of climate production in the Western Sargasso Sea. Deep-Sea Research 6, 235–238.
change on bluefin tuna (Thunnus thynnus) spawning habitat in the Gulf of Salisbury, J.E., Campbell, J.W., Linder, E., Meeker, L.D., Muller-Karger, F.E.,
Mexico. ICES Journal of Marine Science 68, 1051–1062. Vorosmarty, C.J., 2004. On the seasonal correlation of particle fields with
Muhling, B.A., Lamkin, J.T., Richards, W.J., 2012. Decadal-scale responses of larval wind stress and Mississippi discharge in the northern Gulf of Mexico. Dee-Sea
fish assemblages to multiple ecosystem processes in the northern Gulf of Research II 51, 1187–1203.
Mexico. Marine Ecology Progress Series 450, 37–53. http://dx.doi.org/10.3354/ Salmerón-García, O., Zavala-Hidalgo, J., Mateos-Jasso, A., Romero-Centeno, R., 2011.
meps09540. Regionalization of the Gulf of Mexico from space-time chlorophyll-a
Muller-Karger, F.E., 1993. River discharge variability including satellite-observed concentration variability. Ocean Dynamics 61, 439–448. http://dx.doi.org/
plume-dispersal patterns. In: Maul, George (Ed.), Climate Change in the Intra- 10.1007/s10236-010-0368-1.
Americas Sea. United Nations Environmental Programme (Wider Caribbean Sangrà, P., Pascual, A., Rodríguez-Santana, Á., Machín, F., Mason, E., McWilliams, J.C.,
Region) and Intergovernmental Oceanographic Commissions (Caribbean and Pelegrí, J.L., Dong, C., Rubio, A., Arístegui, J., Marrero-Díaz, Á., Hernández-Guerra,
Adjacent Regions). Edward Arnold Publishers, pp. 162–192 (Chapter 8). A., Martínez-Marrero, A., Auladell, M., 2009. The canary eddy corridor: a major
Muller-Karger, F.E., 2000. The spring 1998 NEGOM cold water event: remote pathway for long-lived eddies in the subtropical North Atlantic. Deep-Sea
sensing evidence for upwelling and for eastward advection of Mississippi Water Research 56 (2009), 2100–2114.
(or: How an Errant LC Anticyclone Took the NEGOM for a Spin). Gulf of Mexico Scranton, M.I., Taylor, G.T., Thunell, R., Benitez-Nelson, C.R., Muller-Karger, F.,
Science 1, 55–67. Fanning, K., Lorenzoni, L., Montes, E., Varela, R., Astor, Y., 2014. Interannual and
Muller-Karger, F.E., Walsh, J.J., Evans, R.H., Meyers, M.B., 1991. On the seasonal subdecadal variability in the nutrient geochemistry of the Cariaco Basin.
phytoplankton concentration and sea surface temperature cycles of the Gulf of Oceanography 27 (1).
Mexico as determined by satellites. Journal of Geophysical Research. 96 (C7), Shay, L.K., Goni, G.J., Black, P.G., 2000. Effects of a warm oceanic feature on hurricane
12645–12665. Opal. Monthly Weather Review 128, 1366–1383.
Muller-Karger, F.E., Kavanaugh, M.T., Montes, E., Balch, W.M., Breitbart, M., Chavez, Sheinbaum, J., Candela, J., Badan, A., Ochoa, J., 2002. Flow structure and transport in
F.P., Doney, S.C., Johns, E.M., Letelier, R.M., Lomas, M.W., Sosik, H.M., White, A.E., the Yucatan Channel. Geophysical Research Letters 29 (3). http://dx.doi.org/
2014. A Framework for a Marine Biodiversity Observing Network Within 10.1029/2001GL013990.
Changing Continental Shelf Seascapes. Oceanography 27 (2), 18–23. http:// Shi, C., Nof, D., 1994. The destruction of lenses and generation of wodons. Journal of
dx.doi.org/10.5670/oceanog.2014.56. Physical Oceanography 24 (6), 1120–1136.
NASA CZCS Technical Note, <http://oceancolor.gsfc.nasa.gov/ANALYSIS/PROCTEST/ Siegel, D.A., Behrenfeld, M.J., Maritorena, S., McClain, C.R., Antoine, D., Bailey, S.W.,
ct06m_sr2010.0m/comp_regional_rrs_mission.htm>, 2010. CZCS and SeaWiFS Bontempi, P.S., Boss, E.S., Dierssen, H.M., Doney, S.C., Eplee Jr., R.E., Evans, R.H.,
regional remote sensing reflectance trends.. Feldman, G.C., Fields, E., Franz, B.A., Kuring, N.A., Mengelt, C., Nelson, N.B., Patt,
National Ocean Service, NOAA, <http://stateofthecoast.noaa.gov>, 2011. The Gulf of F.S., Robinson, W.D., Sarmiento, J.L., Swan, C.M., Werdell, P.J., Westberry, T.K.,
Mexico at a Glance: A Second Glance. Washington, DC: U.S. Department of Wilding, J.G., Yoder, J.A., 2013. Regional to global assessments of phytoplankton
Commerce. dynamics from the SeaWiFS mission. Remote Sensing of Environment 135, 77–
Nedbor-Gross, R., Dukhovskoy, D.S., Bourassa, M.A., Morey, S.M., Chassignet, E.P., 91. http://dx.doi.org/10.1016/j.rse.2013.03.025.
2014. Investigation of the relationship between the Yucatan Channel transport Sturges, W., Leben, R., 2000. Frequency of ring separations from the loop current in
and the Loop Current area in a multidecadel numerical simulation. Mar. the Gulf of Mexico: a revised estimate. Journal of Physical Oceanography 30 (7),
Technol. Soc. J. 48 (4), 15–26. 1814–1819.
Nof, D., 2005. The Momentum Imbalance Paradox Revisited. Journal of Physical Sverdrup, H.U.J., 1953. On conditions for the vernal blooming of phytoplankton.
Oceanography 35, 1928–1939. Conseil Perm. Int. Explor. Mer. 18, 287–295.
Nowlin, W.D., Hubertz, J., Reid, R., 1968. A detached eddy in the Gulf of Mexico. Tang, W., Liu, W.T., Stiles, B., Fore, A., 2014. Detection of diurnal cycle of ocean
Journel Marine Research 26, 185–186. surface wind from space-based observations. International Journal of Remote
Nowlin, W.D. Jr., Jochens, A.E., DiMarco, S.F., Reid, R.O., 2000. Physical Sensing 35 (14), 5328–5341. http://dx.doi.org/10.1080/01431161.2014.926413.
oceanography. In: Deepwater Gulf of Mexico Environmental and Trenberth, K.E., Jones, P.D., Ambenje, P., Bojariu, R., Easterling, D., Klein Tank, A.,
Socioeconomic Data Search and Synthesis. Narrative Report. OCS Study MMS Parker, D., Rahimzadeh, F., Renwick, J.A., Rusticucci, M., Soden, B., Zhai, P., 2007.
2000–049. Gulf of Mexico OCS Regional Office. Minerals Management Service. Observations: surface and atmospheric climate change. In: Solomon, S., Qin, D.,
U.S. Department of the Interior, pp. 61–121 (Chapter 4). Manning, M., Chen, Z., Marquis, M., Averyt, K.B., Tignor, M., Miller, H.L. (Eds.),
Oey, L.-Y., 1996. Simulation of mesoscale variability in the Gulf of Mexico: Climate Change 2007: The Physical Science Basis. Contribution of Working
Sensitivity studies, comparison with observations, and trapped wave Group I to the Fourth Assessment Report of the Intergovernmental Panel on
propagation. Journal of Physical Oceanography 26 (2), 145–175. Climate Change. Cambridge University Press, Cambridge, United Kingdom and
O’Reilly, J.E., Maritorena, S., Siegel, D.A., O’Brien, M.C., Toole, D., Chavez, F.P., New York, NY, USA, pp. 235–336.
Strutton, P. Cota, G.F., Hooker, S.B., McClain, C.R., Carder, K.L., Muller-Karger, Tulloch, R.R., Hill, C.C., Jahn, O.O., 2011. Possible spreadings of buoyant plumes and
F.E., Harding, L., Magnuson, A., Phinney, D., Moore, G.F., Aiken, J., Arrigo, K.R., local coastline sensitivities using flow syntheses from 1992 to 2007. Monitoring
Letelier, R., Culver, M., 2000. In: Hooker, S.B., Firestone, E.R. (Eds.), Ocean and Modeling the Deepwater Horizon Oil Spill: A Record-Breaking Enterprise,
Chlorophyll-a Algorithms for SeaWiFS, OC2 and OC4: version 4. SeaWiFS 245–255.
Twilley, R.R., Barron, E.J., Gholz, H.L., Harwell, M.A., Miller, R.L., Reed, D.J., Rose, J.B., Weisberg, R.H., He, R., Liu, Y., Virmani, J.I., 2005. West Florida shelf circulation on
Sieman, E.H., Wetzel, R.G., Zimmerman, R.J., 2001. Confronting climate change synoptic, seasonal, and inter-annual time scales. In: Sturges, W., Lugo-
in the Gulf coast region: Prospects for sustaining our ecological heritage. In: Fernandez, A. (Eds.), Circulation in the Gulf of Mexico. AGU monograph series,
Union of Concerned Scientists, Cambridge. Massachussetts, and Ecological Geophysical Monograph, pp. 325–347.
Society of America, Washington, DC, p. 82. Willis, J.K., Roemmich, D., Cornuelle, B., 2004. Interannual variability in upper ocean
Vidal, V.M.V., Vidal, F.V., Hernández, A.F., Meza, E., Zambrano, L., 1994. Winter water heat content, temperature, and thermosteric expansion on global scales. Journal
mass distributions in the Western Gulf of Mexico affected by a colliding of Geophysical Research 109, C12036. http://dx.doi.org/10.1029/2003JC002260.
anticyclonic ring. Journal of Oceanography 50, 559–588. Wolter, K., Timlin, M.S., 2011. El Niño/Southern Oscillation behaviour since 1871 as
Villanueva, E.E., Mendoza, V.M., Adem, J., 2010. Sea surface temperature and mixed diagnosed in an extended multivariate ENSO index (MEI.ext). Intl. J. Climatol. 31
layer depth changes due to cold-air outbreak in the Gulf fo México. Atmósfera (7), 14. http://dx.doi.org/10.1002/joc.2336.
23 (4), 325–346. Wunsch, C., Heimbach, P., Ponte, R., Fukumori, I., 2009. The global general
Vukovich, F.M., 1988. Loop current boundary variations. Journal of Geophysical circulation of the ocean estimated by the ECCO-consortium. Oceanography
Research 93 (C12), 15585–15591. 22, 88–103.
Vukovich, F.M., 2007. Climatology of ocean features in the Gulf of Mexico using Yoder, J.A., 1979. Effect of temperature on light-limited growth and chemical
satellite remote sensing data. Journal of Physical Oceanography 37 (3), 689– composition of Skeletonema costatum (Bacillariophyceae). Journal of Phycology
707. 15, 362–370.
Vukovich, F.M., 2012. Changes in the loop current’s eddy shedding 854 in the period Young, I.R., Zieger, S., Babanin, A.V., 2011a. Global trends in wind speed and
2001–2010. Int. J. Oceanogr., 18 http://dx.doi.org/10.1155/2012/439042. Article wave height. Science 332 (6028), 451–455. http://dx.doi.org/10.1126/
ID 439042. science.1197219.
Vukovich, F.M., Crissman, B.M., Bushnell, M., King, W.J., 1979. Some aspects of the Young, I.R., Babanin, A.V., Zieger, S., 2011b. Response to comment on ‘‘Global Trends
oceanography of the Gulf of Mexico using satellite and in situ data. Journal of in Wind Speed and Wave Height’’. Science 905. http://dx.doi.org/10.1126/
Geophysical Research 84 (C12), 7749–7768. science.1210548.
Wahl, Thomas, Calafat, Francisco M., Luther, Mark E., 2014. Rapid changes in the Zavala-Hidalgo, J., Morey, S.L., O’Brien, J.J., 2003. Seasonal circulation on the western
seasonal sea level cycle along the US Gulf coast from the late 20th century. shelf of the Gulf of Mexico using a high-resolution numerical model. Journal
Geophysical Research Letters 41. http://dx.doi.org/10.1002/2013GL058777. of Geophysical Research 108 (C12), 3389. http://dx.doi.org/10.1029/
Weisberg, R.H., Black, B.D., Yang, H., 1996. Seasonal modulation of the West Florida 2003JC001879.
continental shelf circulation. Geophysical Research Letters 23 (17), 2247–2250. Zavala-Hidalgo, J., Morey, S.L., O’Brien, J.J., Zambudio, L., 2006. On the loop current
http://dx.doi.org/10.1029/96GL02184. eddy shedding variability. Atmósfera 19 (1), 41–48.
Weisberg, R.H., He, R., Kirkpatrick, G., Muller-Karger, F., Walsh, J.J., 2004. Coastal
ocean circulation influences on remotely sensed optical properties: a West
florida shelf case study. Oceanography 17, 68–75.

Art 1

Uploaded by

Copyright:

Available Formats

Art 1

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Art 1

Uploaded by

Copyright:

Available Formats

Progress in Oceanography 134 (2015) 54–76

Contents lists available at ScienceDirect

Natural variability of surface oceanographic conditions in the offshore

Buoy 42002 Buoy 42003

Buoy 42002 Buoy 42001

F.E. Muller-Karger et al. / Progress in Oceanography 134 (2015) 54–76

question is relevant for assessing impacts of climate change over References

You might also like