The impact of changing climate on phenology, productivity,
and benthic–pelagic coupling in Narragansett Bay
Scott W. Nixon*, Robinson W. Fulweiler, Betty A. Buckley,
Stephen L. Granger, Barbara L. Nowicki, Kelly M. Henry
Graduate School of Oceanography, University of Rhode Island, Narragansett, RI 02882, United States
a r t i c l e i n f o
Received 4 April 2008
Accepted 17 December 2008
Available online 25 December 2008
a b s t r a c t
The timing and magnitude of phytoplankton blooms have changed markedly in Narragansett Bay, RI
(USA) over the last half century. The traditional winter–spring bloom has decreased or, in many years,
disappeared. Relatively short, often intense, diatom blooms have become common in spring, summer,
and fall replacing the summer flagellate blooms of the past. The annual and summer mean abundance
(cell counts) and biomass (chl a) of phytoplankton appear to have decreased based on almost 50 years
of biweekly monitoring by others at a mid bay station. These changes have been related to warming of
the water, especially during winter, and to increased cloudiness. A significant decline in the winter
wind speed may also have played a role. The changes in the phenology of the phytoplankton and the
oligotrophication of the bay appear to have decreased greatly the quantity and (perhaps) quality of the
organic matter being deposited on the bottom of the bay. This decline has resulted in a very much
reduced benthic metabolism as reflected in oxygen uptake, nutrient regeneration, and the magnitude
and direction of the net flux of N2gas. Based on many decades of standard weekly trawls carried out by
the Graduate School of Oceanography, the winter biomass of bottom feeding epibenthic animals has
also declined sharply at the mid bay station. After decades of relatively constant anthropogenic
nitrogen loading (and declining phosphorus loading), the fertilization of the bay will soon be reduced
during May–October due to implementation of advanced wastewater treatment. This is intended to
produce an oligotrophication of the urban Providence River estuary and the Upper Bay. The anticipated
decline in the productivity of the upper bay region will probably decrease summer hypoxia in that area.
However, it may have unanticipated consequences for secondary production in the mid and lower bay
where climate-induced oligotrophication has already much weakened the historically strong benthic–
? 2008 Elsevier Ltd. All rights reserved.
Phenology is the study of relationships between climate and the
regular seasonal progression of biological events, such as the
migration of animals and the flowering of plants. Marine ecologists
have long attended to the arrivals and departures of fish and the
blooming of phytoplankton, but a growing awareness of climate
change has stimulated considerable recent interest in how the
calendar of events in marine ecosystems may be shifting and in
what the ecological consequences of these shifts might be (e.g. Jossi
et al., 2003; Edwards and Richardson, 2004; Stenseth et al., 2004;
Greve et al., 2005; Bering Sea Interagency Working Group, 2006;
Kirby et al., 2007; Pilling et al., 2007; Beaugrand et al., 2008).
It is now clear that the coastal waters off southern New England
have warmed since 1970 by about 1.2?C in their annual mean and
by about 1.7?C in the winter and 1.0?C in the summer (Nixon et al.,
2003, 2004; Oviatt, 2004). In mid Narragansett Bay (RI, USA), the
surfacewaterduring winter (D,J,F) has warmed on averagebyabout
2.2?C since 1960, though there is considerable interannual vari-
ability (Fig. 1). Such warming may have a particularly large impact
on Narragansett Bay and other waters in this region because nearby
Cape Cod forms an important barrier between boreal species to the
north and warm water species to the south (Fig. 2) (Fish, 1925).
While most studies of phenology focus on the responses of
individual species or groups of species to climate change (some
Narragansett Bay examples include Oviatt, 2004; Costello et al.,
2006; Collie et al., 2008), our purpose in this paper is to present and
synthesize some of the evidence that suggests that warming and
associated increased cloudiness and, perhaps, declining wind speed
have had a major impact on productivity, benthic–pelagic coupling,
* Corresponding author.
E-mail address: firstname.lastname@example.org (S.W. Nixon).
Contents lists available at ScienceDirect
Estuarine, Coastal and Shelf Science
journal homepage: www.elsevier.com/locate/ecss
0272-7714/$ – see front matter ? 2008 Elsevier Ltd. All rights reserved.
Estuarine, Coastal and Shelf Science 82 (2009) 1–18
and biogeochemical cycling in Narragansett Bay. We are fortunate
in this effort that Narragansett Bay has been the site of extensive
quantitative phytoplankton studies extending back to the early
1950s (Ferrara,1953; Smayda,1957). Narragansett Bay was also one
of the first and most intensively studied coastal systems in which
benthic oxygen uptake, benthic nutrient regeneration, and benthic
denitrification were measured over annual cycles in the 1970s and
1980s (e.g. Hale, 1974; Nixon et al., 1976, 1980; Seitzinger et al.,
1984). These early studies provide a strong historical base for
comparison with recent conditions (e.g. Fulweiler and Nixon, in
press; Fulweiler et al., 2007).
1.1. A note on Narragansett Bay
Various features of Narragansett Bay have been described
frequently in the literature as part of specialized papers dealing
with particular species or types of organisms or phenomena, but
a detailed scientific overview has not been published for almost 30
years (Kremer and Nixon, 1978). The same is true for a popular
description of the bay’s ecology (Olsen et al.,1980). A set of detailed
reviews of Narragansett Bay phytoplankton, zooplankton, icthyo-
plankton, benthos, demersal fish, organic pollutants, and heavy
metals was prepared in 1989 by Hinga et al., Durbin and Durbin,
Frithsen, Jeffries et al., Quinn, Bender, and Nixon, respectively, but
these are not easily available to readers outside of the area. A new
book devoted to the bay has just appeared, but it is focused on the
history, consequences, and management of nutrient enrichment
(Desbonnet and Costa-Pierce, 2008). For our purposes here, it may
be useful to provide some brief background description of the
ecosystem before discussing changing phenology, productivity, and
benthic–pelagic coupling. Our description is restricted to basic
physical features, phytoplankton, nutrients, and the benthos
because these form the basis of most of the discussion that follows.
1.1.1. Size and shape
The bay as a whole is geographically complex and includes the
Seekonk and Providence River estuaries in the north, a small
shallow side embayment to the west (Greenwich Bay), a much
larger appendage on the east (Mt. Hope Bay), and three longer
passages, the West Passage, the East Passage, and the Sakonnet
Passage (Fig. 2). As a practical matter, Mt. Hope Bay and the
Sakonnet Passage are often not considered as part of Narragansett
Bay proper, a practice we follow here. The Seekonk and Providence
River estuaries are highly urbanized areas that receive most of the
fresh water and sewage thatenters the bay. In contrast with most of
Narragansett Bay proper, they are frequently density stratified with
hypoxic bottomwaters in the shipping channel. Greenwich Bay and
the Upper Bay also experience episodic hypoxia during summer
(e.g. Deacutis et al., 2006; Melrose et al., 2007). The mouth of the
Providence River estuary is marked by Conimicut Pt. (Fig. 2) below
which the bay proper begins.
So defined, Narragansett Bay proper is 234 km2with a mean
depth of 8.7 m at MLW. Mean tidal range is about 1.2 m. The
Providence River estuary is 21.3 km2with a mean depth of 5.2 m.
Distance from Fox Pt. at the head of the Providence River estuary to
the mouth of the bay is about 40 km. A detailed hypsographic
analysis (Chinman and Nixon, 1985) shows that the volume of the
system increases slowly with distance below Fox Pt. until the
divergence of the East and West Passages about 16 km south of
Conimicut Pt., at the top of Prudence Island (Fig. 2). After another
15 km, the depth of the lower East Passage (about 19 m) becomes
much greater than the lower West Passage (about 9 m) and its
volume increases much more rapidly.
1.1.2. Salinity and temperature
Surface fresh water input to the bay averages only about
100 m3s?1(Pilson, 1985) with little additional contribution from
ground water (Nowicki and Gold, 2008; Pilson, 2008). As a result,
salinities are high (32 to about 25 in the bottom water throughout
the system; 32 to about 20 in the surface water, Kremer and Nixon,
1978). The average water residence time in the bay is 26 days
(Pilson, 1985). Water temperatures range from about 1?C to 23?C,
with little difference most of the time between surface and bottom
in the mid and lower bay (Hicks, 1959; Kremer and Nixon, 1978).
Narragansett Bay is now a phytoplankton-based system in
which intertidal wetlands and macrophytes contribute little to the
supply of organic matter. In the late 1800s and early 1900s eelgrass
(Zostera marina) was widely distributed in the upper bay and
Providence River estuary (Nixon et al., 2008), but at present only
y = 0.05x - 88.5
R2 = 0.29
1955 1960 1965 1970 1975 1980 1985 1990 1995 2000 2005 2010
Fig. 1. Mean surface water temperatures during D,J,F in the middle of the West Passage
of Narragansett Bay near Fox Is. (Fig. 2) as recorded by the sampling of D. Pratt, T.J.
Smayda, and the GSO plankton monitoring program (data at http://www.narrbay.org/d_
projects/plankton-tsv/plankton-tsv.htm) and (http://www.gso.uri.edu/phytoplankton/).
Both slope and interceptof the regression are highlysignificant (F¼ 17.212, p ¼0.0002).
Fig. 2. Map of Narragansett Bay showing its location between Cape Cod, MA and Long
Island, NY. Various placed mentioned in text are noted. G. B. is Greenwich Bay. The
solid circle is the site of many historic and recent benthic flux measurements and the
open circle is the long-term demersal fish trawl and phytoplankton monitoring station.
S.W. Nixon et al. / Estuarine, Coastal and Shelf Science 82 (2009) 1–182
some 150 ha are thought to remain, with none above Prudence
Island (Fig. 2) (Bradley et al., 2007). The contribution of benthic
microalgae has not been assessed, but deserves attention in the
shallower parts of the bay. The phytoplankton has been described
by Pratt (1959) and Karentz and Smayda (1998) among others. The
weekly abundance and species composition in the mid West
Passage during recent years can be found online at http://www.gso.
uri.edu/phytoplankton/. There is a strong gradient in surface chlo-
rophyll concentration and primary production from very high
values during summer in the Providence River estuary and Upper
Bay to low at the mouth (Kremer and Nixon, 1978; Oviatt et al.,
2002) (Fig. 3). The up bay–down bay gradients in phytoplankton
dynamics as they relate to nutrient inputs and water clarity have
been analyzed extensively by Smayda and Borkman (2008).
The bay has been intensively fertilized with domestic waste-
water since the introduction of public water supply and sewer
systems beginning in the late 1800s. Detailed reconstruction of
historical and recent nutrient inputs has shown that, until the last
few years, inputs of nitrogen (N) to the bay have remained essen-
tiallyconstant since the early 1970s, and that phosphorus (P) inputs
have declined significantly since the 1970s (Nixon et al., 2008).
Implementation of advanced wastewater treatment has reduced
the direct discharge of dissolved inorganic N (DIN) to the bay from
sewage treatment plants during May–October by about 30% since
2000 (Pryor et al., 2007), but this represents less than a 10% drop in
the annual total N load to the bay. When completely implemented
throughout the watershed and along the shoreline of the bay,
advanced wastewater treatment may ultimately reduce summer N
inputs by 30–40% (Nixon et al., 2008).
Annual mass balances for organic carbon (C), N, and P with full
secondary sewage treatment but prior to the recent move to
advanced wastewater treatment showed that while anthropogenic
sources dominated the inputs of Nand P, phytoplanktonproduction
was by far the larger source of C (Nixon et al.,1995). The size of the
historical winter–spring diatom bloom was regulated by the
availability of DIN and dissolved silica (DSi) (Pratt, 1965; Smayda,
1973; Kremer and Nixon, 1978). During summer primary produc-
tion and phytoplankton standing crops are strongly limited by the
supply of DIN (Oviatt et al., 1995). Concentrations of DIN are very
low (<1–2 mM) in the surface water of the mid and lower West and
East Passages during summer (see nutrient data for a mid bay
station in recent years at the url given above for phytoplankton).
Based on intensive sampling between the mid West Passage and
the head of the Providence River estuary in the mid 1980s, Smayda
and Borkman (2008) concluded that Narragansett Bay could be
divided into three zones, a light limited and Si sensitive ‘‘Enrich-
ment Zone’’ in the Providence Riverestuary, a ‘‘Depuration Zone’’ in
the Upper Bay basin, and a ‘‘Nutrient – Limited, N – Sensitive Zone’’
beginning in the upper West Passage. More recent data seem
consistent with this picture (e.g. Fig. 3).
Over 70% of the bay is deeper than 5 m at MLW (Chinman and
Nixon, 1985), and most of the bottom is believed to be heterotro-
phic, with fine grained silt-clay sediments covering about 65% of
the bottom (Nixon et al., 1995). Almost all of the rest is fine sand.
We say ‘‘believed to be heterotrophic’’ on the basis of vertical light
attenuation and water depth over most of the bay and the rela-
tionships between light and photosynthetic carbon fixation by
microphytobenthos reported by MacIntyre et al. (1996).
Recent biweekly surveys at 16 stations around the bay showed
that the annual mean vertical light attenuation coefficient (?k, m?1)
declined approximately linearly with distance from the head of the
Providence River estuary to the mouth of the bay (Oviatt et al.,
2002) (Fig. 4). Combining these data with the mean depth of
various regions of the bay suggests that on average 1–2% of surface
light reaches the average depth in the Providence River estuary,
Upper Bay, and upper West Passage, 1% reaches the bottom of the
lower West Passage and the upper East Passage, and 0.1% or less
reaches the sediments of the Middle and lower East Passage. Inci-
dent light on the water surface ranges from about 5 to 65 Em?2d?1
over the annual cycle so that sediments at the mean depth of the
various portions of the bay may receive between 0.2 and
40 mmolm?2s?1. Measurements of carbon fixation by micro-
phytobenthos over an annual cycle in Delaware Bay showed very
low rates of photosynthesis at light levels less than about
50 mmolm?2s?1(MacIntyre et al., 1996). We go into this level of
detail because virtually all of the benthic flux measurements made
in the bay have been done in the dark. While microphytobenthos
mayclearly be an important factor influencing benthic fluxes in the
shallow parts of the bay, for the system as a whole their contri-
bution appears to be minimal.
There is a long history of studies describing the benthic
communities in the bay, including Stickney and Stringer (1957),
Phelps (1958), Hale (1974), Frithsen (1984), Rudnick et al. (1985),
Grassle et al. (1985), Levin (1986) and Calabretta and Oviatt (2008).
Frithsen (1989) struggled to synthesize all of the work to that point
to generalize about spatial and temporal trends, but was frustrated
by the range of sieve sizes and collection techniques used, the
common focus on numbers rather than biomass, and the great
seasonal variability in numbers. At least through the 1980s, the
numbers of meiofauna and macrofauna were high in May and June
Mean summer primary
production, g C m-2 d-1
Mean Summer chl a, mg m-3
Distance above (-) or below (+) Conimicut Pt., km
Fig. 3. Mean ?SD summer (J,J,A) near surface chlorophyll (solid points and line) and
below (þ) Conimicut Pt. at the mouth of the Providence River estuary in 1997 (data
from Oviatt et al., 2002, personal communication).
14C particulate primary production with distance above (?) and
-20-100 1020 30 40
Distance above (-) or below Conimicut Pt., km
Fig. 4. Mean annual vertical light attenuation coefficient with distance above (?) and
below (þ) Conimicut Pt. at the mouth of the Providence River estuary in 1997/1998
(data from Oviatt et al., 2002 and personal communication). The gray points are at
stations north of the point where the East and West Passage transects diverge (see
S.W. Nixon et al. / Estuarine, Coastal and Shelf Science 82 (2009) 1–18 3
followed by lowest abundance in late summer and fall, suggesting
food limitation in late summer.
Since much of our discussion of changes in benthic–pelagic
coupling in the bay will focus on the mid bay region just north of
Jamestown or Conanicut Island (Fig. 2), it is worth adding some
detail on the bottom in that area. This is the most intensively and
frequently sampled part of the bay (Frithsen,1989) and it served as
the source of the sediments used in virtually all of the large mes-
ocosm experiments carried out in the Marine Ecosystems Research
Laboratory (MERL) (e.g. Rudnick et al., 1985; Rudnick and Oviatt,
1986; Doering and Oviatt, 1986). The site is 7–8 m deep with
salinity of about 30 that varies little with season. The sediments are
predominantly silt-clay (over 70%) with the remainder largely clay
(almost 20%). Organic contentof the surface sediment was about 4%
with a C/N by weight of 7.6 in the 1970s (Hale, 1974). More recent
analyses suggest that the C/N at the site has increased (mean of
10.9, range 9.5–13.9, Fulweiler, 2007). Light reaching the bottom
may vary between about 4 and 80 mmolm?2s?1. The mid bay
macro benthos is dominated by Mediomastus (a small sub-surface
deposit feeding polychaete) and Nucula (a small bivalve). The
results of recent sampling by C.A. Oviatt and her students can be
seen at: www.gso.uri.edu/merl/data.htm and in Calabretta and
Ovaitt (2008) where they provide complete taxonomic and abun-
dance data. Because of its central location and use in the MERL
experiments, this site has been included in almost all of the benthic
flux measurements made in the bay. Fortunately, it is only about
7 km from the site of the long-term plankton and nutrient moni-
toring site off Fox Island (Fig. 2) studied by Pratt (1959, 1965),
Smayda (1998) and the on-going plankton monitoring cited earlier.
The mid bay region is similar to the average depth of Narragansett
Bay as a whole (8.6 m) and contains the most common sediment
type of the bay. Since it lies about 15–20 km below Conimicut Pt.
(Fig. 2), water column chlorophyll is much lower than in the
Providence Riverestuary but similartothatof the lower bay (Fig. 3).
2. Changing phenology and the winter–spring bloom
It has long been known thaton both the European and American
coasts the most luxuriant diatom growth does not take place in
the warmest months . throughout the shallow waters south of
Cape Cod a rich winter diatom plankton starts usually in
November and continues until March, reaching a maximum in
December (Fish, 1925).
Since the basis for the biogeochemical linkage or coupling of the
benthos and the overlying water begins with the primary produc-
tion of organic matter in the water, we begin our analysis there.
The first reported quantitative study of the phytoplankton in
Narragansett Bay was carried out in the upper West Passage (Fig. 2)
between July 1952 and March 1953 (Ferrara, 1953). The results
showed a strong winter bloom with two lesser blooms during
summer,consistent with the
reported over twenty-five years earlier by Fish (1925). The diatom
Skeletonema costatum was the most abundant and persistent
species, and comprised 80% of the total population. From this
beginning, the ecology of Narragansett Bay has been tightly linked
with the biology and ecology of what has universally been called S.
costatum (Grev.) Cleve. While it is possible that the species was
correctly identified, recent detailed studies of the genus have
shown that S. costatum has, in fact, been frequently and incorrectly
applied to a number of previously known and new species of
Skeletonema (e.g. Sarno et al., 2005). The taxonomic confusion is
understandable – the new distinctions are only apparent to experts
using scanning and transmission electron microscopy and molec-
ular techniques that have only recently become available.
Specimens from Narragansett Bay have been examined using such
techniques, and it appears that at least two species are now
common in the bay, S. japonicum Zingone et Sarno in thewinter and
S. grethae Zingone et Sarno in the summer (Sarno et al., 2005, P.
Hargraves, GSO, URI, personal communication, A. Zingone, Stazione
Kooistra et al., 2008).
Ferrara’s analysis (as well as early studies in the lower bay by
Smayda,1955,1957) was part of a larger and longer sampling study
by David Pratt (1959) that covered 15 stations in the Upper Bay,
Greenwich Bay, and the lengths of the West and East Passages. Pratt
noted the strong winter–spring bloom of diatoms as well as late
summer diatom blooms, ‘‘principally Skeletonema costatum.’’ He
also reported that average standing crops were approximately the
same in the West and East Passages and that, ‘‘the Upper Bay is
more than three times as rich as the lower in both diatoms and
flagellates.’’ The latter observation still seems to hold, at least in
terms of total chlorophyll (Fig. 3).
The best known description of the winter–spring bloom was
provided in a later paper by Pratt (1965, p. 173) based on weekly
samples collected in the upper, mid, and lower West Passage
between January 1959 and June 1963:
‘‘The outstanding feature of the annual cycle is the winter–
spring diatom flowering, which is extraordinary in its time of
inception, intensity, and duration. Logarithmic growth begins
usually in December, and after about a month terminates in
a maximum sometimes exceeding 50,000 cells/ml; this is fol-
lowed by a series of secondary peaks of diminishing amplitude,
and the flowering period ends in late May or June.’’
Unfortunately, the Pratt (1965) paper did not include total cell
counts or data for the mid bay. Fortunately, Pratt did pass on the
total cell count data from the mid bay station to his colleague and
former graduate student, T.J. Smayda, who made them available to
us. Since the mid West Passage station became the location of long-
term plankton studies by Smayda and his graduate students (e.g.
Smayda, 1973; Karentz and Smayda, 1984, 1998; Li and Smayda,
1998; Smayda,1998; Borkman, 2002; Borkman and Smayda, 2009)
is the site of the ongoing monitoring program by the Graduate
phytoplankton/), it is possible to make some direct comparisons
of total cell counts at this station during recent years with the
counts obtained by Pratt and his students a half century ago (Fig. 5).
While the long and intensewinter–spring bloomwas a dramatic
component of the annual cycle in the early 1960s, it is clear that
there were also episodic and intense summer and, in some years,
fall blooms as well. The picture appears very different now, with no
winter–spring bloom in many years, though shorter summer and
fall blooms remain, and these can be quite intense (Fig. 5). The
winters of the three recent years shown in Fig. 5 were all warm
(Fig. 1). The delay and increasingly frequent loss of the winter–
spring diatom bloom appears to have taken place gradually and
erratically over the past 50 years, especially after the mid 1970s
(Fig. 6). This represents a major change in the seasonal cycle of
phytoplankton abundance that had prevailed for at least the
previous 75 years (Fish, 1925).
There appear to have been other important changes as well as
the shift in phenology of blooms. In contrast to the earlier situation,
diatoms now appear to dominate most of the larger blooms
throughout the summer. The major diatom in these summer
blooms is Skeletonema, though Chaetocerous, Leptocylindrus, and
Thalassiosira can also be briefly important (see web site above).
Borkman (2002) reviewed all of the cell count data from the mid
bay station from 1959 to 1996 and found Skeletonema much more
abundant during the summer in recent decades than it was during
S.W. Nixon et al. / Estuarine, Coastal and Shelf Science 82 (2009) 1–184
the 1960s. This appears to represent a change from Pratt’s (1959)
finding that microflagellates were much more abundant than
Skeletonema in spring and early summer during the 1950s (Table 1).
A detailed study of the abundance of different size fractions of the
phytoplankton at this station in 1972–73 by Durbin et al. (1975)
also found flagellates more abundant than diatoms during the
summer. Such a change in the summer phytoplankton community
may have important consequences for secondary production and
biogeochemical cycling. For example, the quahog or hard clam,
Mercenaria mercenaria, appears to feed preferentially on diatoms in
the bay (Gearing et al.,1984) but does not filterat low temperatures
(Loosanoff, 1939). Adult quahogs in the bay are now growing more
rapidly than they did in the 1960s, perhaps in response to more
abundant diatoms in the warmer months when they are feeding
(Henry and Nixon, 2008). The ecological consequences of a decline
in the winter species of Skeletonema (S. japonicum) and an increase
in the summer species (S. grethae) are a rich topic for future
2.1. Declining standing crop
A casual inspection of even the small sample of years in Fig. 5
suggests that the mean abundance of the phytoplankton has also
declined, and a time series plot of the annual mean total cell
numbers over many years is consistent with this impression for
both the annual mean and the winter mean (Fig. 7). A decline in the
abundance of phytoplankton in the mid bay was first reported by
Smayda (1998) based on diatom counts between 1959 and 1980
and by Li and Smayda (1998) based on weekly chlorophyll samples
collected between 1973 and 1990. They noted a decline of
2.25 mgm?3over the 18-year period. If additional chlorophyll data
106 cells l-1
106 cells l-1
106 cells l-1
Fig. 5. Comparison of total phytoplankton cell counts at the long-term phytoplankton monitoring station in the middle West Passage (Fig. 2) during the early 1960s and early 2000s.
Historic data of Pratt (1965) provided courtesy of T.J. Smayda (see Dwyer, 1980); recent data from the GSO plankton monitoring url given in text (courtesy of P. Hargraves). Earlier
data pooled surface, mid depth, and near bottom; recent data mean of near surface and near bottom.
Fig. 6. Time of maximum bloom development in the West Passage of Narragansett Bay
based on various sources, including Pratt (1965), Smayda (1998), Li and Smayda (1998),
MERL monitoring (C. Oviatt, personal communication), and the GSO plankton
Seasonal mean concentrations (103cells per liter) of phytoplankton sampled weekly
in the surface water at station W8 off Fox Island in the mid West Passage during
1955–1956 (from Pratt, 1959).
Winter–Early SpringLate Spring–Early Summer Late SummerFall
S.W. Nixon et al. / Estuarine, Coastal and Shelf Science 82 (2009) 1–185
collected by Smayda between 1991 and 1996 (personal communi-
cation) are added to the earlier Li and Smayda (1998) time series
along with the measurements made more recently by the GSO
plankton monitoring program (see web site above), it appears that
the decline continued through 2006 (Fig. 8). While the frequent
decline or loss of the winter–spring bloom is particularly conspic-
uous (as noted by Li and Smayda, 1998; Smayda, 1998), there has
also been a marked decline in the mean summer chlorophyll
(historical data provided by T.J. Smayda, more recent data from the
GSO phytoplankton monitoring) (Fig. 8). And there has been
a marked decline in the overall abundance of Skeletonema. Bork-
man’s detailed study (2002) concluded that, ‘‘Winter–spring Skel-
etonema bloom duration declined from ca. six weeks in 1959–1963
to three weeks in 1978–1982 while first quarter abundance
declined from 6000 cellsml?1(1959–1963) to ca. 1200 cellsml?1
2.2. Declining primary production
Such changes in the abundance and standing crop of the
phytoplankton raise the question of changes in the primary
production of the bay. Unfortunately, the question is impossible to
answer firmly with direct measurements because there have been
only a few studies of primary production and they have used quite
different techniques (e.g. Vargo, 1979; Oviatt et al., 1981 used
oxygen changes; Furnas et al., 1976; Oviatt et al., 2002 used14C
uptake; and Durbin and Durbin, 1981 used laboratory derived
assimilation numbers with field chlorophyll). Different studies have
also reported theirresultsin different ways (e.g.depth integrated or
per unit volume) and only the Oviatt et al. (1981, 2002) studies
attempted to cover the entire bay. Perhaps the most direct
comparison is between the14C based estimates of Furnas et al.
(1976) and Oviatt et al. (2002) near the mid West Passage station.
The former made measurements during 1974 (310 gCm?2y?1) and
the latter during 1997–1998, a year with no winter–spring bloom
(160 gCm?2y?1). Extensive studies by Keller (1988) have shown
that14C uptake by the phytoplankton in Narragansett Bay (as in
many other coastal systems reviewed by Brush et al., 2002) can be
calculated very closely using a simple BZI approach (biomass as chl
a, photic depth, incident light). Using this approach, the decline in
chlorophyll in the mid bay (Fig. 8) suggests that primary production
there may have declined from about 370 gCm?2y?1in the mid
1970s to about 210 gCm?2y?1in 2005 (Fulweiler et al., 2007). Both
comparisons suggest a reduction in primary production at mid bay
of about 40–50% over 25 years. As we argue in a later section on
benthic fluxes, even this large decline may be an underestimate.
A decline in the supply of organic matter to an ecosystem is
called oligotrophication (Nixon, in press), the opposite of the well-
known eutrophication that has been occurring due to increasing
nutrient enrichment in many other coastal systems around the
world (e.g. Nixon, 1995; Rabalais and Nixon, 2002; Valiela, 2006).
While some of the ecological consequences of oligotrophication
have become clear in lakes (e.g. Nay, 1996; Anderson et al., 2005),
the phenomenon is only just beginning to be recognized in marine
systems (Ozaki et al., 2004; Carstensen et al., 2006; Greening and
Janicli, 2006; Philippart et al., 2007; Nixon, in press). While
management interventions have been responsible for the few cases
cited, we hypothesize that the oligotrophication of Narragansett
Bay (at least to this point) has been induced by climate change. The
only other case of climate-induced oligotrophication of a marine
ecosystem that we are aware of is the remarkable indication of
declining carrying capacity in the Bering Sea based on stable
isotopic analysis of carbon in whale baleen (Schell, 2000) as well as
declines in benthic macro-infauna biomass and benthic oxygen
uptake in the northern Bering Sea (Grebmeier et al., 2006).
In spite of the impressive cell counts and chlorophyll of the
winter–spring bloom, most of the primary production by the
phytoplankton occurred during summer and early fall in the early
1970s (Vargo,1979). Furnas et al. (1976) estimated that almost 75%
of the total annual14C fixation during 1974 took place between May
and September, inclusive, and over 40% occurred in July and August
alone. Summer production exceeded winter in the more recent
measurements as well (Oviatt et al., 2002). As shown by Furnas
et al. (1976), most of this summer productionwas supported by DIN
recycled within the water column, though benthic fluxes of NH4
could have provided 20–50% of the net mid bay phytoplankton N
demand in July and August of 1974. As we show later, that benthic
contribution is now very much reduced, as is the summer standing
crop of phytoplankton (Fig. 8).
2.3. Declining zooplankton?
While the focus of this paper is on the impact of changes in the
phytoplankton on benthic–pelagic coupling, a question naturally
arises about the responses of the zooplankton. Answering this
of zooplankton studies in Narragansett Baywas reviewed byDurbin
and Durbin (1988), who concluded that it was virtually impossible
to document trends in abundance up to that point because of the
wide variety of sampling techniques that had been used in the
different studies. Second, the warming winters (Fig. 1) have also
Cells, 106 L-1
Fig. 7. Mean annual and winter total phytoplankton cell counts at the long-term
monitoring site in the middle West Passage of Narragansett Bay (Fig. 2). Data sources
as in Fig. 6 plus T.J. Smayda (personal communication). Both the slope (p <0.05) and
intercept (p<0.05) for the annual mean vs. year are significant. The slope (p <0.0002)
and the intercept (p <0.0002) are also significant for the winter mean.
Chl a, mg m-3
Fig. 8. Mean annual and summer chlorophyll at the long-term monitoring site in the
middle West Passage of Narragansett Bay (Fig. 2). Annual data from Li and Smayda
(1998) through 1990. Historical summer data and annual data from 1991 to 97 from
T.J. Smayda (personal communication). All data from 1999 to 2006 from the GSO
plankton monitoring url given in text. The high annual mean of 1995 was associated
with a very cold winter (mean D,J,F ¼1.7?C, see Fig. 1). The mean annual chlorophyll
decline is significant if two outliers (1995 and 1996) are removed (p <0.0001). The
summer chlorophyll decline is significant (p <0.05). Expanded from Fulweiler and
Nixon (in press).
S.W. Nixon et al. / Estuarine, Coastal and Shelf Science 82 (2009) 1–186
changed the phenology of the comb jelly, Mnemiopsis leidyi, a major
predator on the summer zooplankton. As a result the Mnemiopsis
now overwinter in greater numbers, proliferate much more quickly
in the spring, and reach greater abundances in the summer than
they did in previous decades (Sullivan et al., 2001). Costello et al.
(2006) have shown a dramatic decrease in the maximum abun-
dance of Acartia tonsa, the dominant summer copepod in the bay,
when comparing the mean values obtained in various studies
between 1951 and 1983 with their own study in 2000–2003. But
that comparison must also suffer to some degree from sampling
differences and, in any case, cannot distinguish between top down
and bottom up influences. Even in the 1970s, when chlorophyll
levels were considerably higher, both food limitation and predation
appeared to be important in limiting A. tonsa abundance and
production in this system (Durbin and Durbin, 1981). In analyzing
samples collected during two years in the mid 1980s, Smayda and
Borkman (2008) found positive spatial correlations between mean
annual chlorophyll and mean annual zooplankton dry weight
biomass that would suggest there has probably been a significant
decline associated with the reduction in phytoplankton.
2.4. Declining organic deposition
It is difficult to quantify how the changes in the timing and
magnitude of phytoplankton abundance and production have
altered the deposition of organic matter on the bottom, including
its quality as well as its quantity. Sediment trap measurements are
not useful for this purpose because of resuspended sediments
(Oviatt and Nixon, 1975). At least two attempts have been made to
get around this problem by using the large (13 m3, 5 m deep) MERL
mesocosm tanks in which both14C labeling of phytoplankton and
sediment traps can be used (Rudnick and Oviatt, 1986; Riebesell,
1989, respectively). Unfortunately, the results vary widely. In any
case, extrapolation of these results to the bay is questionable. The
tanks used in both studies were run in ‘‘batch mode’’ without any
exchange of water with the adjacent Narragansett Bay as was the
more common mode of operation of the facility. As noted by Rie-
besell (1989), this meant that the size and duration of the bloom
were limited by the initial stock of nutrients in the mesocosm.
We can make an indirect assessment of maximum potential
deposition during the traditional winter–spring bloom using a bay-
wide survey of nutrient concentrations and chlorophyll made
during an annual cycle in 1972–1973 (Kremer and Nixon,1978). The
bloom took up virtually all of the DIN in the bay as it developed
from late January to early March and DIN concentrations below
Conimicut Pt. remained near zero from March through June. With
an initial DIN concentration of about 15 mM (and mean depth of
8.6 m) and assuming Redfield organic matter, this uptake would
have been associated with new primary production of about
10 g Cm?2. To hold the concentration of DIN near zero, the phyto-
plankton would also have to have taken up virtually all of the DIN
entering the system. Our measurements of DIN in the rivers and
sewage treatment plants entering the bay during the 1970s (Nixon
et al., 2005, 2008) amount to approximately 625 mmolm?2in
February through June, inclusive. This requires additional new
production of about 50 gCm?2for a total of 60 gCm?2. This is still
an underestimate of the total new production because it does not
include the carbon fixed using the DIN in Rhode Island Sound water
brought into the bay by the tides and entrained in the gravitational
circulation. The sub pycnocline water in the sound has higher DIN
concentrations than the mid and lower bay surface water. This
input of DIN is much more uncertain, but various estimates have
arrived at about 1.1 mmolm?2d?1(Nixon and Pilson, 1984; Nixon
et al., 1995; Chaves, 2004). Over the five months this suggests new
production of some 13 gCm?2, giving a total of 73 gCm?2. Of
course, not all of this production was deposited on the bottom –
some was respired in the water column and some must have been
flushed fromthe system. But it does point toan upperconstraint for
the traditional winter–spring bloom deposition.
Studies in the MERL mesocosms and in natural systems inwhich
sediment traps can be used effectively have shown that the depo-
sition of organic matter from blooms is greatly reduced when the
water is warmer (e.g. Smetacek, 1984; Rudnick and Oviatt, 1986;
Keller et al.,1999). While grazing by zooplankton consumed little of
the traditional winter–spring bloom in Narragansett Bay (Vargo,
1976; Deason, 1980) or in MERL mesocosms with a winter–spring
bloom (Keller and Riebesell, 1989), the current situation with
almost all of the blooms coming during spring, summer, and fall
must be associated with greatly reduced deposition of organic
matter on the bottom of the bay. Moreover, the cold water bloom
deposition was almost completely made up of sinking phyto-
plankton (Keller and Riebesell, 1989) with a low C/N, while the
material deposited during warmer conditions is of lower nutri-
tional value with a higher C/N (Smetacek, 1984). Studies of the
chemical composition of Skeletonema by Yoder (1979) also showed
higher N/chl a and C/chl a in cells grown at 0?C than in cells grown
at higher temperatures (5,10,16, 22?C). As concluded by Smetacek
(1984, p. 533), ‘‘.the seasonal cycle in new and regenerated
production in the pelagic system is of vital importance to the
benthos both in terms of quantity and quality of the food supply.’’
3. Consequences for the benthos
Numerous studies have documented that during a relatively
brief period the spring phytoplankton bloom in temper-
ate.regions can deliveras muchas half of the total annual input
of organic carbon to the benthos.Earlier blooms may occur in
colder water, which would reduce consumption by pelagic
heterotrophs and result in the input of a greater proportion of
planktonic production to the bottom sediments.
Townsend and Cammen (1988)
The dramatic shift of the Narragansett Bay winter–spring
diatom bloom to spring, summer, and/or fall (Figs. 5 and 6) is an
extreme example of the up to 20 day shifts in spring bloom initi-
ation that Townsend and Cammen estimated using a decade of
incident light data for Sheepscot Bay, Maine. The thesis of their now
classic paper was that there might be a significant decline in food
available for juvenile demersal fish in years when the spring bloom
was delayed by a greater frequency of cloudy days. Later blooms
would occur in warmer water where more of the primary
production would be consumed by zooplankton. As a result, less of
the organic matter would sink and support secondary production
by benthic infauna, the important food for demersal fish.
While Narragansett Bay might seem a perfect setting to test this
hypothesis, the data on benthic infauna are still confounded by all
of the factors that frustrated Frithsen (1989). The major problem
with the more recent sampling (which uses 300 mm mesh screens)
and most of the historical studies is thatonly numbers of organisms
have been reported, and these vary dramatically with time of
sampling – even a week or two can make a big difference (Grassle
et al.,1985; Frithsen,1989). Efforts are underway by C.A. Oviatt and
her students to estimate benthic biomass, but even with such
estimates in hand there are few (if any) comparable data from the
past for documenting long-term changes.
3.1. Declining epibenthic biomass in winter
There are consistent records of demersal fish and other epi-
benthic animal abundances at the mid bay long-term plankton
S.W. Nixon et al. / Estuarine, Coastal and Shelf Science 82 (2009) 1–187
monitoring station (Fig. 2) that extend back to 1959. This standard
weekly trawl sampling was begun by C.J. Fish (whose observations
on the winter diatom bloom we quoted earlier). It was maintained
for decades by H.P. Jeffries (e.g. Jeffries and Terceiro, 1985) and is
now part of the GSO bay monitoring program overseen by J.S. Collie
(www.gso.uri.edu/fishtrawl/). The RI Department of Environmental
Management (DEM) has also monitored demersal fish during
spring and fall at about 26 stations in the bay since 1979. Based on
these data sets, there is no question that the abundance of demersal
finfish in the bay has declined dramatically (at least during winter)
and that the mean size of fish has declined significantly (e.g. Oviatt
et al., 2003; Oviatt, 2004; Collie et al., 2008). However, it is difficult
toknow if, and towhat degree, the change in timing and magnitude
of the phytoplankton production has contributed to the decline in
the fish. The decline is common in non-commercial as well as
commercial benthic species (Oviatt et al., 2003), but it is particu-
larly evident in winter flounder (Pseudopleuronectes americanus),
a heavily exploited species. Using the DEM data, Oviatt et al. (2003)
estimated that the biomass of demersal fish (including northern sea
robin, red hake, skate, dogfish, tautog, windowpane flounder, and
winter flounder) declined by about 75% between 1980–1985 and
1995–2000. As those authors point out, the warming of the bay
itself may have contributed to the decline of boreal species such as
the winter flounder. It is also possible that the warming may have
had an indirect negative influence specifically on winter flounder
by advancing the time when the sand shrimp, Crangon septemspi-
nosa, a major predator on post-settlement winter flounder,
becomes active (Jeffries, 2001; Taylor, 2003). But it is also possible
that food limitation has played a role. A recent study of juvenile
winter flounder in Narragansett Bay noted that there has been little
recovery of stocks here despite a closure of the bay fishery and
suggested that ‘‘.reduced size at age could be one factor respon-
sible for the increased mortality of winter flounder juveniles.’’
(Buckley et al., 2008).
The interpretation of the decline in demersal finfish is also
complicated by large increases in the numbers of decapods in the
bay, especially cancer crabs (Cancer irroratus and C. borealis) and
lobster (Homarus americanus). After making an admittedly rough
conversion from numbers to fresh weight for the decapods, Oviatt
et al. (2003) and Oviatt (2004) estimated that their increase in total
biomass (including shell) was about half the decline in finfish
biomass. Since shell accounts for about 50–75% of the fresh weight
of crabs and lobsters while skeletal material is usually less than 5%
of the fresh weight of fish (Vinogradov, 1953), the meat biomass of
the decapods may only amount to some 12–25% of the 75% loss in
demersal finfish biomass.
We have estimated the average total fresh meat biomass of
organisms captured during winter (J,F,M) and summer (J,A,S) in the
standard 30 min GSO weekly trawl sampling at mid bay between
1960 and 2005 (Fig. 9). Since finfish biomass data have only been
collected since 1994, we applied the average weight per individual
of each species during each month of the last eleven years to the
count data prior to 1994. Because the size of the individual fish has
been declining (Collie et al., 2008) the decline in biomass is almost
certainly greater than that shown in Fig. 9. For the decapods, we
assumed that live meat biomass was 37% of the fresh weight of
crabs and lobsters and 50% of the fresh weight of starfish and
conchs (which are occasionally caught in large numbers in the
trawls; C. Oviatt, personal communication and Sisson, 1973,
respectively). While there is a great deal of interannual variability,
winter biomass is now down by 80% while there is no significant
trend in the summer biomass. The summer fish, however, are
largely composed of pelagic feeders such as butterfish, scup, and
squid while the winter contains many more benthic feeders. Collie
et al. (2008) have noted a pronounced switch from benthic to
pelagic fish in the bay since 1980. The food required to support the
demersal finfish in the past may also have been even greater than
a simple comparison of the total epibenthic standing crop suggests
since the production (and ingestion) per unit biomass of finfish is
generally higher than for the invertebrates (Valiela, 1995, p. 238)
that are now more abundant.
It would be surprising if the oligotrophication of Narragansett
Bay had not played some role in the decline of demersal fish. Cross
system comparisons have shown strong correlations between
primary production and fisheries yield (Nixon, 1988), primary
production and fish production (Iverson, 1990), and fish yield and
chlorophyll concentrations (Ware and Thomson, 2005) in marine
ecosystems. Moreover, the link between declining primary
production and declining fish abundance and production has
frequently been demonstrated in lakes (e.g. Nay, 1996). Neverthe-
less, we must admit that the situation with regard to the demersal
fish may be too convoluted to offer definitive evidence of a benthic
response (at least in part) to the changing phenology of the
phytoplankton blooms. For that we must go to something simpler –
the changing fluxes of oxygenand nutrients and nitrogengas across
the sediment–water interface.
3.2. Changes in benthic–pelagic fluxes
3.2.1. Historical background
The first of what became a great many measurements of the
uptake of dissolved oxygen and the net flux of ammonium, nitrate
plus nitrite, phosphate, and silicate between the bottom of Narra-
gansett Bay and the overlying water were made in July,1973 (Hale,
1975). That work grew into the first measurements of these fluxes
over a complete annual cycle in a marine system and focused on
three stations along the West Passage, including a mid bay station
off the north end of Conanicut Island, not far from the long-term
plankton monitoring station (Fig. 2) (Nixon et al.,1976; Kremer and
y = -0.451x + 906.4
R2 = 0.38
Biomass, kg tow-1
1977 1987 19972007
Biomass, kg tow-1
Fig. 9. Estimated mean live biomass (excluding shell) of demersal epibenthic animals
collected by the GSO long-term fish trawl (approximately weekly sampling) off Fox Is.
in the middle West Passage of Narragansett Bay during winter (J,F,M) and summer
(J,J,A) (see Fig. 2). Seasons defined by taxonomic composition of the fish. See text for
assumptions. Counts provided by J. Collie (personal communication) from the GSO
demersal fish monitoring program. Sampling details are given in Collie et al. (2008).
There is no significant trend in the summer fish but the overall decline in the winter is
highly significant (F¼ 20.02; p<0.0001). The decline since 1983 is also significant
(p¼0.021) while there was not a significant slope between 1970 and 1983.
S.W. Nixon et al. / Estuarine, Coastal and Shelf Science 82 (2009) 1–188
Nixon, 1978; Nixon et al., 1980). While the first study used large
(2400 cm2, 27 l) in situ chambers, subsequent work employed
a variety of smaller and differently shaped cores collected by divers
and incubated in the laboratory (e.g. McCaffrey et al., 1980; Kelly,
1983; O’Reilly,1984; Nowicki and Nixon,1985). As noted earlier, the
number of flux measurements involving mid bay sediments
increased dramatically because the area became the collection site
for sediments used in virtuallyall the MERL experiments during the
1970s and 1980s (e.g. Pilson et al., 1980; Kelly and Nixon, 1984;
Nixon et al.,1984; Oviatt et al.,1984; Kellyet al.,1985; Doering et al.,
1986; Rudnick and Oviatt, 1986; Keller et al., 1987). Benthic
metabolism measurements at MERL normally used a large incu-
bation chamber that covered the entire 2.5 m2sediment tray in
each mesocosm (Oviatt, 1981). Comparisons of fluxes measured in
various ways did not show consistent differences, but there was
a surprising lack of a gradient in fluxes measured from the Upper
Bay to the mouth (Hale, 1975; Nixon et al., 1976).
3.2.2. Oxygen uptake and nutrient regeneration
The benthic oxygen uptake measurements of the mid 1970s
showed that about 140 gCm?2y?1was being respired by the
benthos or some 45% of the primary production reported by Fur-
nas et al. (1976) for the mid bay in 1974. This seems a large frac-
tion relative to a number of other productive coastal systems
reviewed by Nixon (1981) and by Hopkinson and Smith (2005) in
which correlations were found between the autochthonous and
allocthonous carbon supply and benthic respiration. Such corre-
lations suggest that primary production may have been on the
order of 450–500 gCm?2y?1in the 1970s (Table 2). Phyto-
plankton production estimates for the mid bay using weekly
chlorophyll measurements and laboratory determined assimila-
tion ratios (Durbin et al., 1975; Durbin and Durbin, 1981; Table 2)
also suggest that Furnas et al. (1976) may have underestimated
carbon fixation, perhaps because they used 24 h long incubations.
When DIN was virtually depleted during the highly productive
summer months (e.g. Kremer and Nixon, 1978), their samples may
have been strongly nutrient limited. If so, then the decline in
primary production in the mid bay has been even greater than we
Benthic metabolism was also strongly influenced by water
temperature, though there was a marked seasonal effect such that
rates of oxygen uptake and nutrient regenerationwere significantly
higher in the spring than they were at equivalent temperatures in
the fall. This was interpreted as a sign of food limitation of the
benthos during late summer and fall – as the water warmed in the
spring, the benthos had an abundant supply of organic matter from
the deposition of the winter–spring bloom, while little of the high
summer production reached the bottom (Nixon et al., 1976, 1980;
Kelly and Nixon, 1984). The coupling of the water column and the
benthos through organic matter deposition and subsequent
nutrient regenerationwas clearlya major feature of the metabolism
and biogeochemical cycling of the bay in the 1970s and early 1980s
(Nixon et al.,1976; Kremer and Nixon,1978; Kelly and Nixon,1984).
3.2.3. Net N2fluxes
Narragansett Bay was also the site of the first direct measure-
ments of the net flux of N2gas across the sediment–water interface
using intact, unamended cores collected over an annual cycle
(Seitzinger et al., 1980, 1984; Seitzinger, 1982). Stations were
sampled in 1978–1979 and included the lower Providence River
estuary, the mid bay where all the benthic fluxes of dissolved
oxygenand nutrients were made byothers, and justoutside the bay
in Rhode Island Sound. The results showed that, ‘‘Denitrification
represents a major sink for fixed N in the bay; annually the N2
production is equal to about 50% of the fixed N loading to the bay
from rivers, land, and sewage’’ (Seitzinger et al., 1984, p. 73).
Nowicki (1994) modified the water-replacement, gas chromatog-
raphy technique used in the early work and returned tothe mid bay
site in 1986. Her results showed a lower average rate of denitrifi-
cation (385 mmolm?2y?1compared with 520 mmolm?2y?1), but
it was not clear if this apparent decline was real or an artifact of the
use of different methods. In either case, denitrification in the
sediments was still an important process in the bay.
3.2.4. The new situation-dramatic reductions in benthic respiration
and nutrient regeneration; the appearance of N fixation
The gas chromatograph analyses used by Seitzingerand Nowicki
have now been replaced by much more precise measurements of
changes in the N2/Ar ratio in water incubated over the sediments
using membrane inlet mass spectrometry, MIMS (Kana et al.,1994,
1998; Eyre et al., 2002; Gardner et al., 2006). It was the develop-
ment of this new technique that motivated us to begin making
benthic flux measurements again, including N2, in the summer of
2005, after a gap of almost twenty years.
Oxygen uptake rates calculated from the first cores collected in
2005 at the mid bay site were remarkably low compared with the
measurements of the 1970s and 1980s. So were the rates of
ammonium release and phosphate release. These low rates per-
sisted and were significantly (p<0.01) different from the earlier
measurements (Fulweiler and Nixon, in press) (Fig. 10). Additional
measurements in 2006 continued to show much lower benthic
metabolism (Fig.10). The exchange of nitrate (not shown) remained
low and erratic with both uptake and release as observed in the
past (Nixon et al., 1976; Fulweiler, 2007). An interesting exception
to the marked decline in benthic fluxes was the release of dissolved
silica, which remained at historic rates (Fig. 10).
Because of the strong link between primary production in the
overlying water and the metabolism of the sediments below
(Nixon, 1981; Kelly and Nixon, 1984; Seitzinger and Giblin, 1996;
Boynton and Kemp, 2000; Hopkinson and Smith, 2005) it is
compelling to attribute these major changes to the decline of
phytoplankton biomass and production. But the change in the
timing of blooms may be equally (or perhaps even more) impor-
tant. As Smetacek (1984), Townsend and Cammen (1988), and
others pointed out decades ago, the shift from cold water to warm
water blooms as the major source of new organic matter in the bay
almost certainly means that less of the new production now rea-
ches the benthos. Moreover, as discussed earlier, the nutritional
quality of the material deposited during summer and fall, when
dissolved inorganic nitrogen is virtually depleted and much of the
production is supported by recycled nutrients, is probably much
lower (high C/N) than material deposited from a winter bloom. We
must say ‘‘probably’’ because for reasons discussed earlier, it is
difficult, if not impossible, to make direct measurements on the
freshly sedimenting material in this system. As Smetacek (1984, p.
Various estimates of primary production in mid Narragansett Bay during the 1970s.
24 h14C incubations
Regession of benthic O2uptake against
water column primary productionþorganic
C input in many marine systemsa
As above using regression of Hopkinson
and Smith (2005)
Using mean annual assimilation ratio and field
As above, but for March –September only
Furnas et al. (1976)
This study 510
Durbin et al. (1975)
aMeasured annual O2uptake¼140 gCm?2y?1assuming RQ of 1 (Nixon et al.,
1976); allocthonous C input¼75 gCm?2y?1(Nixon et al., 1995).
S.W. Nixon et al. / Estuarine, Coastal and Shelf Science 82 (2009) 1–189
533) put it, ‘‘The particles sinking out of such [regenerating]
systems are truly wastes, i.e. they are composed of refractory
material with lowessential element content.’’ We noted earlier that
the C/N content of the surface (0–2 cm) sediments at the mid bay
site appears to have increased since the 1970s. The lack of change in
the benthic silica flux is consistent with this interpretation. The
warm water pelagic grazers and decomposers have no nutritional
need for silica, which falls to the bottom in fecal pellets and/or
diatom debris. The return of dissolved silica from the sediments
thus remains high. Carbon, nitrogen, and phosphorus, however, are
assimilated, respired, mineralized, and recycled in the water
column with only a small fraction reaching the benthos.
The N2flux measurements that motivated our return to benthic
flux work were even more surprising. In 2005, the mean annual
rate of denitrification at the mid baysitewas significantly (p<0.01)
lower than the measurements a quarter of a century earlier
(40 mmolm?2h?1compared to 75 mmolm?2h?1) (Fulweiler and
Nixon, in press). Of course, the methods were quite different so it
was hard to know if this apparent change was real. That became
clear in the summer of 2006, when we found that bacteria in the
sediments were actually fixing large amounts of nitrogen (Ful-
weiler et al., 2007). This was completely unexpected since deep
heterotrophic marine sediments had never been shown to fix N at
any significant rate. In fact, a15N tracer study using mid Narra-
gansett Baysediments in the mid 1980s showed virtually no or only
very low N fixation (Seitzinger and Garber, 1987). With the
exception of sea grass rhizospheres, mangroves, and salt marshes,
even shallow marine sediments have shown no or very low rates of
N fixation (see reviews by Capone, 1983; Howarth et al., 1988).
A recent exception, however, is work in Texas estuaries by Gardner
et al. (2006) and McCarthy et al. (2008) using MIMSthat found high
rates of net N fixation at some locations. Since virtually all of the
earlier work on N fixation used the acetylene reduction technique
(seldom, if ever, calibrated with15N), it is possible that more direct
measurements of net N2 flux using MIMS will discover that N
fixation in heterotrophic marine sediments is more common and
important than previously believed.
The switch from net denitrification to net N fixation in the
sediments appears tobe another consequence of the long decline in
the quantity and/or quality of organic input to the bottom of
Narragansett Bay. Experimental enrichment of N fixing sediments
with spray dried phytoplankton to mimic historical boom deposi-
tions quickly reversed the net N2flux and returned the sediments
to net denitrification (Fulweiler et al., 2007, 2008). While the
evidence from measurements made in the summer and fall of 2006
suggests that net N2fixation may be confined to warmer temper-
atures, the summer rates of fixation were sufficiently high (about
200 mmolm?2h?1at the mid bay site) that they made the sedi-
ments a net source of N to the bay over the annual cycle amounting
to 20–60% of the direct discharge of N to the bay from sewage
treatment plants (Fulweiler et al., 2007). The contrast with the
situation described by Seitzingeret al. (1984) that wequoted earlier
is dramatic. It remains to be seen if net N fixation will come to
characterize the sediments of the bay or if the net N2flux varies
fromyear-to-year or on shorter time scales with bloom dynamics in
the overlying water. January and February of 2008 were marked by
a dramatic winter phytoplankton bloom of historic proportions and
summer 2008 benthic flux measurements are eagerly anticipated.
3.2.5. A weaker coupling
With hindsight, we should have anticipated that the major
changes in thewatercolumnwould have produced largechanges in
the metabolism of the benthos (as speculated by Keller et al.,1999;
Oviatt et al., 2002 for the benthic macro-infauna). But the truth is
that until we saw the changes in benthic fluxes we did not look
hard at the water column. And whenwe did, two pieces of evidence
came together for the first time. The decline in phytoplankton first
reported for cell counts by Smayda (1998) and for chlorophyll by Li
and Smayda(1998) coveredthe periods 1959–1980 and 1973–1990,
respectively. While Smayda had continued his weekly monitoring
through 1996, the results were not yet generally available. The GSO
weekly plankton monitoring program did not begin until 1999.
While these data were available on the web site, they had not been
put together with the earlier published values. Once we put the
05 101520 25
mg m-2 h-1
05 10 1520 25
µ µmol m-2 h-1
µ µmol m-2 h-1
5 10 152025
µ µmol m-2 h-1
Fig.10. Net fluxes of dissolved oxygen, ammonium, phosphate, and dissolved silicate across the sediment–water interface in mid Narragansett Bay (see Fig. 2) during the 1970s and
1980s (gray circles) and in 2005 and 2006. Historical data from various sources cited in text and author’s unpublished measurements. Oxygen, ammonium, and phosphate fluxes
modified and expanded from Fulweiler and Nixon (in press).
S.W. Nixon et al. / Estuarine, Coastal and Shelf Science 82 (2009) 1–1810
GSO data together with the Li and Smayda (1998) paper, the full
extent of the decline became clear (Fulweiler and Nixon, in press;
Fulweileret al., 2007). Now, through the courtesy of T.J. Smayda, we
have also been able to include his results for 1991–1996 as well as
a breakout of his summer data to be married with the GSO summer
data (Fig. 8). We have also been able to assemble the trend in cell
counts (Figs. 5 and 7). While there are some procedural differences
between the way chlorophyll samples were handled in the Smayda
laboratory and in the GSO monitoring program (use of a 1 atm.
manifold pump to filter samples vs 50 ml syringes; 5–10 ml sample
vs 25 ml sample; goal of 1 month maximum storage of frozen
(?20?C) filters vs up to 3 or 4 months on some occasions), it seems
unlikely that these would have produced the continuation of the
decline (D. Borkman, Smayda Laboratory, GSO, personal commu-
nication). And the cell counts should be free of such concerns.
Taken together, the shift in the phenology of phytoplankton
blooms, the declining phytoplankton chlorophyll and cell counts at
the mid baystation, the marked decline in benthic metabolism, and
the switch of net N2gas flux across the sediment–water interface
that is reversible with organic enrichment provide compelling
evidence that benthic–pelagic coupling has changed dramatically
in Narragansett Bay. The decline in epibenthic animal biomass
during winter is also consistent with this evidence. It seems a great
irony that the bay in which so much of the early work was done
documenting the strong coupling between the benthos and the
water column should now be the place where we see some of the
first signs that such coupling may be fundamentally altered.
4. What caused the changes?
The discussion and speculation that follow should be tempered
with the perspective that we do not yet have a complete and
convincing mechanistic explanation for the timing of the tradi-
tional winter–spring bloom, despite decades of research. Smayda
(1973, p. 219) made a detailed study of the 1971–1972 bloom and
concluded that, ‘‘Temperature, light, nutrients, grazing, possibly
species interactions, and hydrographic disturbances regulated the
seasonal dynamics of Skeletonema costatum.’’ A quarter of a century
later, and now established as one of the most experienced and
respected marine phytoplankton ecologists, he would write
(Smayda,1998, p. 563), ‘‘Phytoplankton variability is commonplace
and easily detectable, but the limited data on its properties of scale,
multiple varieties, different periods, amplitudes and response
times, and biological interactions make quantification of cause and
effect and predictions problematic.’’ All the complexities that made
understanding the historical bloom such a challenge remain to
make us humble about our ability to understand its passing.
Perhaps it is as Smetacek and Pollehne (1986, p. 425) concluded
long ago, ‘‘Rather than repeating the same cycle rigidly each year,
Nature improvises as she goes along.’’
4.1. The MERL experiment and the Narragansett Bay model
There are at least two prevailing and related theories, and both
may be at least partially correct. The first is that the warming we
described in the introduction, especially the winter warming, has
increased grazing pressure such that the winter–spring bloom
cannot develop, at least during particularly warmyears. Oviatt et al.
(2002) have shown a weak (R2¼0.32) inverse correlation between
mean weekly chlorophyll over the month of the winter–spring
bloom and the mean water temperature during January–March,
inclusive. More direct evidence for the importance of grazing
comes from a mesocosm experiment using the MERL tanks (Keller
et al., 1999). In that study, triplicate tanks were warmed or cooled
a few degrees above or below the 1977–1989 Narragansett Bay
mean winter temperature such that the average temperature
during the December through January experiment was 5.3 or 1.3?C
in the warm and cool tanks, respectively. There were some
complications caused by the settlement of large numbers of blue
mussels, Mytilus edulis, in two of the warm and one of the cool
tanks, but the two other cool tanks developed winter phyto-
plankton blooms while the warm tanks did not and zooplankton
biomass was greater in the warmed systems. They also found (p.
351) that, ‘‘. warm systems. supplied little freshly sedimented
material to the benthos’’ while cool systems showed ‘‘.sedimen-
tation of large amounts of fresh phytodetritus to the benthos.’’
While the MERL results seem clear, there are two points that are
relevant to these results as a complete explanation for events in the
bay. First, the 4?C temperature difference between cool and warm
treatments may seem small, but all but one of the 26 winters in the
Oviatt et al. (2002) regression fall within this range. In other words,
the treatments represented close to the extreme range in winter
temperature conditions in the bay over recent decades (Fig. 1).
Second, the competing (or perhaps complementary) theory is that
lower light in warmer winters (due to increased cloudiness) is an
important contributing factor to delays or losses of the winter–
spring bloom (Borkman, 2002) (Fig. 11). In this context it may be
important that the apparent vertical light attenuation coefficient
(?k, m?1) of the water in the MERL mesocosms is much higher than
in the bay. This is a consequence of the water being contained in an
opaque cylinder. The mean k value did not differ between treat-
ments in the warming experiment, but it averaged 1.22 m?1(Keller
et al., 1999), far greater than average k values anywhere in the bay
(Fig. 4). A reasonable question is whether or not the phytoplankton
that could not keep up with the grazing pressure in the warmed
MERL tanks might have been able to keep up and even bloom if
they had been exposed to more light. The importance of light in
initiating the traditional winter–spring bloom in Narragansett Bay
has been well established by incubation experiments (Smayda,
1973), analyses of time series field data at the mid bay station
(Hitchcock and Smayda,1977; Nixon et al.,1979) and by microcosm
experiments and numerical modeling (Kremer and Nixon, 1978;
Nixon et al., 1979).
winter–spring bloom in Narragansett Bay is commonly traced to
Martin (1966, 1970). But what Martin (1966, p. 67) concluded was,
‘‘Grazing severely limited the standing crop of this diatom
[Skeletonema] when primary production was stopped or slowed
by inadequate light intensities .’’
y = -6.0852x + 150.2
R2 = 0.2
Fig. 11. Irradiance measured by the Eppley Laboratory in Newport, RI and surface
water temperature at the long-term plankton monitoring station in the middle West
Passage of Narragansett bay (see Fig. 2). Data from 1959 to 1996 from http://www.
narrbay.org/d_projects/plankton-tsv/plankton-tsv.htm. Data for 1999 through 2005
from http://www.gso.uri.edu/phytoplankton/ and the Eppley Laboratory. Values are
the mean of weekly temperature measurements and hourly light measurements
during January through April. Both the intercept (p<0.0001) and the slope
(p¼ 0.0057) are significant (F¼ 8.51, p< 0.0057).
S.W. Nixon et al. / Estuarine, Coastal and Shelf Science 82 (2009) 1–1811
In other words, zooplankton grazing might be important in
terminating the bloom, but not in preventing its initiation if light
and nutrients were adequate. In fact, an experimental study of
Skeletonema growth at low temperatures showed that maximum
growth (i.e. with optimum light and nutrients) increased expo-
nentially with temperature between 0 and 10?C at 12.6% per?C,
equivalent to a Q10in this temperature range of 3.5 (Yoder, 1979).
This contrasts with the Q10of 1.9 between 4 and 16?C for instan-
taneous gut clearance by the dominant winter zooplankton, Acartia
hudsonica (Wlodarczyk et al.,1992). Since the copepods cannot feed
any faster than they can clear their guts, this comparison suggests
that if light and nutrients are abundant, Skeletonema (presumably S.
japonicum) will respond to winter warming more positively than
the copepods. Based on additional field and laboratory studies,
Vargo (1976, p. 118) reached essentially the same conclusion as
Martin, ‘‘The impact of grazing in regulating the diatom, and
specifically the Skeletonema population in Narragansett Bay was
An extreme example of the temperature vs light question with
regard to the winter–spring bloom was also explored numerically
in the Narragansett Bay ecosystem model (Kremer and Nixon,
1978). Perhaps the strongest features of the model were the elab-
orate and detailed treatments of the influence of light on phyto-
plankton growth and the influence of temperature on zooplankton
feeding, egg hatching, and the growth rate and development time
of juveniles (followed as daily cohorts). One numerical experiment
was to eliminate the annual water temperature cycle and run the
model as it was normally done (the standard run), but with
temperature set at the annual mean of 11.5?C throughout the year.
When this was done, the winter–spring bloom still developed,
though a bit earlier than in the standard run. Overall, however, the
winter–spring bloom was little altered even by a major winter
warming when light was adequate. The reciprocal experiment in
which the standard seasonal temperature cycle was used but light
was held constant at various levels showed clearly that the timing
and magnitude of the winter phytoplankton biomass was very
sensitive to light.
4.1.1. Changes in the wind?
While Borkman’s (2002) observation that warmer winters tend
to be cloudier fits nicely with the field and model analyses doc-
umenting the importance of light in initiating the traditional
winter–spring bloom, the data shown in Fig. 11 are not completely
convincing as a full explanation for the delay and/or elimination of
the bloom. A recent analysis of mean wind speed adjacent to the
bay during the three windiest months of the year (F,M,A) has
shown a much more dramatic decline in recent decades, from
about 19 kmh?1in the late 1960s and early 1970s to about
15.5 kmh?1in the early 2000s (Pilson, 2008). Since the work of the
wind in vertical mixing varies roughly as the cube of the speed
(Niiler and Kraus, 1977; Husby and Nelson, 1982), this represents
about a 45% decline in mixing potential. While the mid and lower
bay are usually described as well mixed, especially during winter,
such description is based on the fact that there are no or only small
differences in the concentrations of various substances (e.g. nutri-
ents, salt, chlorophyll) between near surface and near bottom
waters. But the daily history of light exposure experienced by the
phytoplankton in such a ‘‘well mixed’’ water column might vary
a great deal depending on the wind.
The earlier analyses of field data for the winters of 1972–1973
and 1975 (Hitchcock and Smayda, 1977; Nixon et al., 1979,
respectively) found that delayed cloudy winter blooms beganwhen
the meanlightinthe water
(4.2 Jm?2d?1), the same as found by Riley (1967) in his studies of
the annual bloom in Long Island Sound. In a rapidly mixed water
column exceeded40 Lyd?1
column, the mean light may be a good description of the light field
seen by an average phytoplankton cell. But with slower mixing,
perhaps a higher incident light may be required to produce a strong
bloom. Thus, there may be an interaction between the decreasing
surface light due to increasing cloud cover in warm winters and the
decrease in vertical mixing due to declining winter wind speeds.
The truth is that we do not really know. As noted by Steemann-
Nielsen and Hansen (1959), ‘‘It must be made clear at once that in
Nature the investigation of the dependence of plankton photo-
synthesis on light is an extremely intricate problem.’’ More recent
evidence that this is still the case is provided by a series of
instructive numerical experiments carried out by Lucas et al. (1998)
in which the complex interactions of light and mixing on phyto-
plankton blooms were simulated for south San Francisco Bay, an
admittedly very different system in terms of stratification.
4.1.2. Benthic grazers
As the absence of a winter bloom in the cool MERL tank with
a dense settlement of blue mussels showed, zooplankton are not
the only potential winter grazers in the bay. And the impact of
benthic filter feeders is not limited by the relatively long lag time of
zooplankton population responses (e.g. Petersen, 2004). The
activity of many benthic filter feeders declines or ceases at low
temperatures (Gerritsen et al., 1994). For example, the amount of
time that the most widespread bivalve in the bay, the quahog or
hard clam, Mercenaria mercenaria, spends open and filtering
declines sharply below 10?C (Loosanoff, 1939; Kremer and Nixon,
1978). While the blue mussel and the soft clam (Mya arenaria, also
present in parts of the bay) do continue filtering at rates that
decline approximately linearly with declining temperature, at least
down to 2?C (Riisgård, 2004), there is no evidence that their
abundance (or that of other suspension feeders) has been
increasing regularlyover recent decades. On the otherhand, it must
be recognized that there has been no regular monitoring of such
populations in the bay during this time. The depth of the bay may
also be sufficient that benthic grazers are unlikely to be responsible
for the loss of the winter-spring bloom. This was the conclusion of
a detailed modeling study of benthic grazing in (much warmer)
Chesapeake Bay for parts of the system that averaged 9 m,
essentially the same as the mean depth of Narragansett Bay
(Gerritsen et al., 1994).
4.1.3. Light, temperature, and nutrient limitation
It is possible that light limitation, warmer temperatures, and
enhanced grazing have worked in concert to produce the changes
that have taken place. Both DIN and DSi reach their maximum
concentrations in late fall-early winter (Fig. 12). The longer the
bloom is delayed by inadequate light, the more DIN and DSi are
flushed offshore where concentrations are lower. Until 2004, when
advanced wastewater treatment began to be implemented at some
treatment plants between May–October, inclusive, the input of DIN
to the bay from sewage treatment plants was relatively constant
throughout the year. The same is true of direct atmospheric
deposition. But the input from rivers varies strongly with river flow
(Nixon et al., 1995, 2008). For DSi, of course, river flow is the only
significant source. A simple steady-state conservative mixing
model using average flushing rates as a function of fresh water
inflow (Pilson, 1985) suggests that in the absence of biological
uptake, volume-weighted average DIN concentrations in the bay as
a whole would range from about 15 mM in January and February to
less than 10 mM in all remaining months. Thus the amount of new
production supported by the nutrient draw down is reduced by
a third for later blooms. In addition, the amount of DIN brought into
the bay by rivers also declines from almost 6 mmolm?2d?1in
January and February to about 4.6 mmolm?2d?1in March,
S.W. Nixon et al. / Estuarine, Coastal and Shelf Science 82 (2009) 1–1812
3.7–3.9 mmolm?2d?1in April and May, and 2.7–2.9 mmolm?2d?1
in June and July. The combined effect is that nutrient limitation will
set in earlier for later blooms, thus exacerbating the impact of
higher grazing rates at warmer temperatures.
The loss of the winter–spring bloom deposition appears to
reduce the mean standing crop of summer phytoplankton as well
(Fig. 8), perhaps because the nutrients formerly stored on the
bottom are now flushed offshore. As a result, the fluxof ammonium
from the sediments is much reduced during summer (Fig. 10) and
the summer primary production may have become even more
dependent on rapid recycling within the water column than it was
in the 1970s (Furnas et al., 1976). Interestingly, the mean concen-
trations of DIN and DSi in the mid bay during summer appear little
changed from conditions during the early 1970s, while reductions
in DIP in rivers and sewage treatment plant effluents (Nixon et al.,
2008) are reflected in significantly lower DIP concentrations
(Table 3). Production in the mid and lower bay during summer was,
and remains, strongly nutrient limited. We don’t know if the
reduction in ammonium fluxes from the sediments is reflected
directly in the reduced standing crop of phytoplankton or if the C/N
of the phytoplankton has also changed.
The overall impact of the changing phenology on seasonal
nutrient cycles also remains unclear. Writing of the 1950s and
1960s, Pratt (1965, p. 181) noted that, ‘‘.silicate is usually unde-
tectable during most of the winter and spring and reappears as
flowering wanes, attains an annual maximum (often in July).’’ The
situation is often very different now, with DSi remaining high
during much of the winter as in 2005 and 2006 and low during
much of the summer (Fig. 12). Some other changes in nutrients
probably have little or nothing to do with changes in the ecology of
the bay. For example, Pratt (1965) reported that the fall nitrate
maximum in the late 1950s and early 1960s seldom exceeded 5 mM
while it reached 5–10 mM in mid West Passage by the early 1970s
and remains at that level during the fall maximum today. This was
almost certainly due to improvements in sewage treatment that
raised oxygen levels in the rivers draining to the bay and thus
stimulated nitrification of the formerly high ammonium levels in
the rivers (Nixon et al., 2008). Phosphate has declined as a result of
P removal from detergents and improved sewage treatment (Nixon
et al., 2008).
Close comparison of the time course of phytoplankton cell
counts, chlorophyll, DIN and DSi at the mid bay station reveals
periods of match and mismatch. For example, the winter of 2003/
2004 showed sharp declines in DIN and DSi in late December/early
January that coincided with rapid increases in chlorophyll and cell
counts – the historical pattern (Fig.12, Arrow A). During the winter
of 2005, however, the decline in DIN and DSi was slower, did not
Concentration (µ µM)
Cell Count (106 cells l-1)
Chlorophyll-a (mg l-1)
Fig. 12. Approximately biweekly measurements of mean near surface and near bottom chlorophyll a, total phytoplankton cell counts, dissolved inorganic nitrogen (DIN), and
dissolved silicate (DSi) at the long-term plankton measuring station in the middle West Passage of Narragansett Bay from June 2003 through June 2006. Plankton data from http://
www.gso.uri.edu/phytoplankton/. Nutrient data are from our laboratory and available at the same url.
Comparison of mean surface water nutrient concentrations (mM?SD) during
summer (J,J,A) in Narragansett Bay near the middle of the West Passage (off Fox
Island) in 1972–1973 and in 2003 through 2006. Early data from Kremer and Nixon
(St.8) (1978), more recent data from our laboratory http://www.gso.uri.edu/
S.W. Nixon et al. / Estuarine, Coastal and Shelf Science 82 (2009) 1–1813
take place until late Januaryand February, and was accompanied by
only a modest increase in chlorophyll compared with cell counts
(Fig. 12, Arrow B). This was repeated in winter 2006 when cell
counts increased strongly with nutrient declines in late winter but
chlorophyll showed little response (Fig. 12, Arrow C). The stoichi-
ometry of the chlorophyll response is also of concern, though we
recognize that N:C:Chl ratios are quite variable depending on
species composition, nutritional status, etc. For example, a 15 mM
drop in DIN due to phytoplankton uptake should support about
1200 mg of C fixation. Based on the extensive review by Cloern et al.
(1995), over 80% of C:Chl ratios of light-limited phytoplankton fall
between 20 and 60 (Brush et al., 2002), thus suggesting a chloro-
phyll increase of 20–60 mgm?3compared with the observed
increases of 2–6 mgm?3in 2005 and 2006 (Fig.12). Of course, some
of the chlorophyll is lost to grazing, sinking, and flushing, but the
discrepancy is still disconcerting. The magnitude of the cell count
increases associated with the DIN uptake seems more in line with
expectations. Assuming that the blooms were largely Skeletonema,
fixation of 1200 mg of C might yield an increase of about 13 million
cells per liter (Parsons and Takahashi, 1973, p. 39), a response
roughly consistent with the observations (Fig. 12).
4.1.4. The speed of benthic responses
Regardless of what is ultimately shown to have been the cau-
se(s) for the remarkable changes observed in Narragansett Bay, the
response time of benthic–pelagic coupling has probably been
rapid. While14C labeled organic additions in the MERL mesocosms
have shown that there may be time lags of several months
between deposition and remineralization of the winter–spring
bloom (Rudnick and Oviatt, 1986), annual mass balance calcula-
tions have shown that only a small amount (perhaps 5–10%) of the
primary production and allocthonous organic carbon accumulates
in the sediments over an annual cycle (Nixon et al., 1995). Shorter
term studies of the remineralization of15N labeled organic matter
in sediment cores have suggested that the half life of detrital N
landing on the sediment surface is on the order of 1–2 months in
the spring (8?C) and 2–3 weeks in the fall (16?C) (Garber, 1982).
suggested that 95% of added organic matter would be remineral-
ized within 4–8 months at 10?C. Additional deposition experi-
ments with cores at 15?C showed that over 80% of organic matter
deposited on the bottom was remineralized over an annual cycle
and that benthic metabolism responded very rapidly (a few days)
to organic additions at 15?C (Kelly and Nixon, 1984). Taken
together with the experiments showing that the net flux of N2gas
across the sediment–water interface can switch from net denitri-
fication in one summer to net fixation in the next, and that organic
additions can reverse the net flux from fixation to denitrification
in a matter of a few days (Fulweiler et al., 2007, 2008), this body of
evidence suggests that the strength of benthic–pelagic coupling in
the bay may vary markedly from year to year and, perhaps, from
month to month.
14C labeled deposition studies by O’Reilly (1984) also
4.1.5. A legacy effect?
The situation in Narragansett Bay may be particularly inter-
esting with regard to the net N2flux between the water and the
sediments because of the bay’s long historyof intensive fertilization
with anthropogenic nitrogen. As an urban estuary, the bay has been
a highly productive phytoplankton-based system for well over
a century (Nixon et al., 2008). While the prehistoric bay may also
have been highlyproductive, much of that productionwas probably
provided by sea grasses, kelp beds, and epibenthic algae in the
shallows (Nixon, 1997; Nixon et al., 2008). The distribution and
accumulation of organic matter in the sediments from these sour-
ces would have been quite different from the pattern that emerged
with the switch to dense phytoplankton blooms in the upper bay in
the late 1800s (Nixon, 1989, Fig. 3). If the changing climate and
changing phenology of blooms has meant that less good quality
(lower C/N) organic matter has been reaching the sediments during
the past few decades, the long legacy of relatively high C/N organic
matter accumulated in the sediments may now provide a substrate
inwhich N fixing bacteria are at an advantage overdenitrifiers. If so,
the bay may now be in a delicate balance – in cold, clear winters
a rich winter–spring bloom results in a large amount of high quality
organic matter reaching the bottom, thus stimulating benthic
secondary production, sustaining higher summer nutrient regen-
eration, higher new production by the summer phytoplankton, and
some nitrogen removal in the sediments. Perhaps this also happens
following some intense blooms in the summer associated with
episodic storm inputs of nitrogen. On the other hand, increasingly
warm and cloudy winters maylead toreduced new production, less
organic deposition, and, perhaps, net N fixation by the sediments in
summer. It is too soon to know if the latter is a recurrent
5. What next?
The great question for scientists and managers alike is what will
happen over the next few years as a large, politically mandated
reduction in N inputs to Narragansett Bay from sewage treatment
plants during warmer months is fully implemented. Since the city
of Providence, the largest source of sewage N at the head of the bay,
will also be treating its storm water runoff for N removal between
May and October, even episodic DIN inputs will be reduced during
the warm months. How will this management intervention interact
with the changing phenology of the bay? The climate-induced
oligotrophication of the mid bay may have gone as far as it can go,
at least in terms of new production during years with warm, cloudy
winters (Figs. 7 and 8). Most of the summer DIN input from land
appears to have been taken up in the Providence River estuary and
the Upper Bay even when anthropogenic inputs were high (Fig. 3,
Kremer and Nixon,1978; Oviatt et al., 2002; Smayda and Borkman,
2008). In the future virtually none of the DIN from land may reach
the West and East Passages. If there is little or no winter–spring
storage of N on the bottom of the mid and lower bay, summer new
production there may be largely a function of DIN inputs from
Rhode Island Sound. While this is an inexhaustible source, DIN
concentrations in surface water there are also very low during
summer, so the net input must be small. Total
probably continue to be supported by rapid recycling in the water
column, but food for the benthos may be in very short supply.
14C uptake will
5.1. Linking upper bay primary production and lower bay
The reduction of N in sewage discharged to the rivers draining
into the bay and directly tothe upper bay will almost certainly have
the intended effect of reducing primary production and algal
biomass in the estuaries and Upper Bay and probably reduce
summer hypoxia in those areas. But what will be the impact of that
loss of primary production on the rest of the bay? How important
has the anthropogenically stimulated primary production at the
head of the bay become to secondary production, especially the
production of benthic animals, in the rest of Narragansett Bay?
Recent extensive measurements of the stable isotopes of N (15N and
14N) in hard clams throughout the bay showed that the animals all
appear tobe living in largepartonphytoplanktonwho’s growth has
been supported by heavy anthropogenic N in the Providence River
estuary and Upper Bay (Oczkowski et al., 2008). While N taken up
in the estuary and Upper Bay might support the growth of the
S.W. Nixon et al. / Estuarine, Coastal and Shelf Science 82 (2009) 1–1814
phytoplankton as they are carried some way down the bay as
shown by York et al. (2007) in Waquoit Bay on Cape Cod, it is also
possible that phytoplankton sinking out of the surface waters in the
upper bay might be delivered to the benthos of the mid and lower
bay. For those used to thinking of classic estuarine circulation this
might seem surprising. However, recent numerical modeling and
observational ADCP data have shown that under different wind
conditions there is a net flow of bottomwater from the upper tothe
lower bay through either the West or the East Passage (Rogers,
2008). This may begin to explain why age specific growth rates of
the hard clams have not uniformly changed throughout the bay
from rates measured half a century ago despite the decline in mid
bay phytoplankton (Henry and Nixon, 2008). While growth rates of
juvenile clams (<2 years) have decreased, adults actually appear to
have been able to take advantage of the warming temperatures
and/or changing phenology (more diatoms in summer) to increase
their growth rates (Henry and Nixon, 2008). If this food source is
significantly reduced and the transport of DIN from the upper bay
to the mid and lower bay is virtually eliminated, most of the
secondary production in the bay will have to be supported by the
already much reduced new production in the mid and lower bay.
An optimistic prediction might be that summer N2fixation will
add enough N to the mid and lower bay that new production there
will be enhanced while hypoxiawill be improved in the much more
stratified Upper Bay and Providence River estuary. Given the large
storage of high C/N organic matter in the bay’s sediments, high
rates of net N2fixation might continue for a long time. It might
become a future much like the distant past, with clearer waters in
the shallower upper parts of the bay and estuary allowing some
regrowth of eelgrass and more epibenthic algae. N2 fixation in
eelgrass beds and by photosynthetic bacteria in the shallower
sediments might be complemented with heterotrophic N2fixation
in the deeper carbon-rich sediments. The biologically fixed N might
ultimately be excreted into the overlying water to support phyto-
plankton production in the mid and lower bay. We might create
a bay without the steep gradients in phytoplankton primary
production and chlorophyll that have probably characterized the
past century or more, but with equal or higher overall secondary
production. The best of both worlds. On the other hand, the marine
N cycle in general, and in Narragansett Bay in particular, has proven
tobe morecomplicated and interestingthanwethought it was only
a decade ago. And the biogeochemical cycling of this critical
limiting nutrient in this bay is embedded in an ecosystem that has
been changing dramatically in response to changing climate and
other anthropogenic pressures. We should expect further surprises
and watch closely.
5.2. Scaling up
It is reasonable to ask if other estuaries and bays in this area are
experiencing changes similar to those observed here. Summer
primary productivity and phytoplankton blooms have declined in
nearby Boston Harbor, but that is due to large-scale sewage diver-
sion (Libby et al., 2007). Benthic metabolism has also declined in
the harbor (Tucker et al., 2008). More intriguing are widespread
declines in the abundance of diatoms (but not total phytoplankton)
and copepods in Massachusetts Bay and Cape Cod Bay from 1992
through 2006 (Libby et al., 2007). However, there appears to have
been no statistically significant change in primary production by
the phytoplankton in Massachusetts Bay during that time (Oviatt
et al., 2007), and benthic oxygen uptake and nutrient regeneration
(with the exception of DSi) have not declined (Tucker et al., 2008).
South of Cape Cod, water column chlorophyll declined significantly
between 1987 and at least 1998 in Buzzards Bay, about 45 km east
of Narragansett Bay (Turner et al. in press Phytoplankton
chlorophyll has also declined markedly in Peconic Bay at the
eastern end of Long Island, but this appears to be a response to
a decline in nitrogen concentrations (Suffolk County Department of
Health Services, 2007). Narragansett Bay is unique in this region in
the long time series of phytoplankton data and benthic metabolism
measurements and in its relative stability of nutrient inputs during
recent decades (Nixon et al., 2008). These circumstances may have
come together to allow us to observe a development that would not
be apparent in many other coastal areas. After an extensive review
of phytoplankton bloom dynamics in coastal marine systems
around the world, Cloern and Jassby (2008) concluded, ‘‘.our
synthesis shows that the timing and amplitude of phytoplankton
variability usuallydo not follow
synchronous seasonal patterns in nearshore coastal waters where
variability tied to large-scale climate can be overwhelmed by local
Those of us who study Narragansett Bayare not used to drawing
parallels between our small, shallow, urban estuary and much
larger continental shelf systems like the Bering Sea or the North
Sea. But it is almost eerie to read the account of changes taking
place in the northern Bering Sea given by Grebmeier et al. (2006).
They, too, report warming waters, a reduced flux of carbon to the
bottom, declining benthic oxygen uptake rates, declining benthic
populations, an increase in pelagic compared to demersal fish, and
an overall weakening of what was ‘‘tight pelagic–benthic coupling’’
in the late 1980s. In the North Sea, too, recent work has shown that
climate-induced changes in the benthos can propagate through
pelagic food chains from phytoplankton to fish and fundamentally
alter benthic–pelagic coupling (Kirby et al., 2007). In fact, there is
a renewed appreciation for the ‘‘tight trophic coupling’’ demon-
strated in the pelagic ecosystems of the Northeast Atlantic that may
make them particularly sensitive to climate changes (Richardson
and Schoeman, 2004). A recent review of data from chlorophyll
through copepods to cod in the North Atlantic found that the
greatest susceptibility to warming appeared to be in areas where
the mean annual temperature lay between 9 and 12?C, with
a maximum between 9 and 10?C (Beaugrand et al., 2008). The
coastal waters off southern New England south of Cape Cod aver-
aged about 10.5?C from the late 1800s until recent warming from
the 1970s and reached 11.9?C during the 1990s (Nixon et al., 2004).
Of course, there are important differences and complications
in comparing these much larger systems with Narragansett Bay –
for example, the loss of sea ice is probably not an important
factor here and predation by diving sea ducks and marine
mammals has not so far attracted much attention. But the
atmospheric and hydrographic forcings which seem to be
responsible for dramatic ecosystem changes in the northern
Bering Sea and the north Atlantic may, like the smaller scale
changes here, be ultimately linked to global warming and its
associated climate changes. That such large changes in marine
ecosystems of such a wide range of scale may be due to secular
changes that appear small relative to interannual variation is
a sobering observation and an important caution.
recurrent and spatially
We acknowledge and thank our colleagues T.J. Smayda, H.P.
Jeffries, and C.A. Oviatt whose dedication to the long-term study
andmonitoringof Narragansett Baymade this reviewand synthesis
possible. Discussions with many others at the Graduate School of
Oceanography and elsewhere have also informed our interpreta-
tion of the data presented and of the bay in general, including
M.E.Q. Pilson, C. Kincaid, J. Frithsen, P. Hargraves, J. Collie, L. Harris,
R. Chinman, M. Brush, D. Borkman, A. Durbin, D. Durbin, J. Kelly,
A. Keller, A. Oczkowski, P. DiMilla, P. Doering, D. Rudnick, J. Chaves,
S.W. Nixon et al. / Estuarine, Coastal and Shelf Science 82 (2009) 1–18 15
K. Hyde, S. Hale, J. Garber, S. Seitzinger, J. Kremer, P. Kremer,
B. Sullivan, V. Lee, C. Calabretta, and many others. L. Lucas and
J. Cloern provided helpful comment and analysis on the possible
importance of wind and tidal mixing on the light climate in the bay.
We thank A. Oczkowski for help with illustrations and some
statistical analyses and L. Harris for application of the BZI model to
chlorophyll data. Comments from Ivan Valiela and an anonymous
reviewer improved the manuscript. We also thank Ivan Valiela for
the invitation to prepare this review.
Anderson, N., Jeppesen, E., Soendergaard, M., 2005. Ecological effects of reduced
nutrient loading (oligotrophication) on lakes: an introduction. Freshwater
Biology 50, 1589–1593.
Bering Sea Interagency Working Group, 2006. Climate Change and the Bering Sea
30 pp. Available from: http://www.afsc.noaa.gov/Publications/ProcRpt06.htm.
Beaugrand, G., Edwards, M., Brander, K., Luczak, C., Ibanez, F., 2008. Causes and
projections of abrupt climate-driven ecosystem shifts in the North Atlantic.
Ecology Letters 11, 1157–1168.
Borkman, D., 2002. Analysis and Simulation of Skeletonema costatum (Grev.) Clev
Annual Abundance Patterns in Lower Narragansett Bay 1959 to 1996. Ph.D.
thesis. Graduate School of Oceanography, University of Rhode Island, Narra-
gansett, RI, 395 pp.
Borkman, D., Smayda, T., 2009. Multidecadal (1959–1997) changes in Skeletonema
abundance and seasonal bloom patterns in Narragansett Bay, Rhode Island,
USA. Journal of Sea Research 61, 84–94.
Boynton, W., Kemp, M., 2000. Influence of river flow and nutrient loads on selected
ecosystem processes – a synthesis of Chesapeake Bay data. In: Hobbie, J. (Ed.),
Estuarine Science-a Synthetic Approach to Research and Practice. Island Press,
Washington, DC, pp. 269–298.
Bradley, M., Raposa, K., Tuxbury, S., 2007. Report on the Analysis of True Color Aerial
Photography to Map and Inventory Zostera marina L. in Narragansett Bay and
Block Island Sound. University of Rhode Island Environmental Data Center,
Narragansett Bay National Estuarine Research Reserve, Save the Bay, 16 pp.
Brush, M., Brawley, J., Nixon, S., Kremer, J., 2002. Modeling phytoplankton
production: problems with the Eppley curve and an empirical alternative.
Marine Ecology Progress Series 238, 31–45.
Buckley, L., Collie, J., Kaplan, A., Crivello, J., 2008. Winter flounder larval genetic
population structure in Narragansett Bay, RI: recruitment to juvenile young-
of-the-year. Estuaries and Coasts 31, 745–754.
Calabretta, C., Ovaitt, C., 2008. The response of benthic macrofauna to anthropo-
genic stress in Narragansett Bay, Rhode Island: a review of human stressors and
assessment of community conditions. Marine Pollution Bulletin 56, 1680–1695.
Capone, D., 1983. Benthic nitrogen fixation. In: Carpenter, E., Capone, D. (Eds.),
Nitrogen in the Marine Environment. Academic Press, New York, pp. 105–137.
trend reversal: a Danish case study. Limnology and Oceanography 51, 398–408.
Chaves, J., 2004. Potential Use of
Narragansett Bay. Ph.D. thesis. Graduate School of Oceanography, University of
Rhode Island, Narragansett, RI, 113 pp.
Chinman, R., Nixon, S., 1985. Depth–Area–Volume Relationships in Narragansett
Bay. RI Sea Grant Marine Technical Report 87. Graduate School of Oceanography,
University of Rhode Island, Narragansett, RI, 64 pp.
Cloern, J., Jassby, A., 2008. Complex seasonal patterns of primary producers at the
land–sea interface. Available from: Ecology Letters http://www3.interscience.
Cloern, J., Grenz, C., Vidergar-Lucas, L., 1995. An empirical model of the Phyto-
plankton chlorophyll:carbon ratio – the conversion factor between productivity
and growth rate. Limnology and Oceanography 40, 1313–1321.
Collie, J., Wood, A., Jeffries, H., 2008. Long-term data reveal climate forcing of coastal
fish community structure. Canadian Journal of Fisheries and Aquatic Sciences
Costello, J., Sullivan, B., Gifford, D., 2006. A physical–biological interaction under-
lying variable phonological responses to climate change by coastal zooplankton.
Journal of Plankton Research 28, 1099–1105.
Deacutis, C., Murray, D., Prell, W., Saarman, E., Korhun, L., 2006. Hypoxia in the
upper half of Narragansett Bay, RI, during August 2001 and 2002. Northeastern
Naturalist 13, 173–198.
Deason, E., 1980. Grazing of Acartia hudsonica (A. clausi) on Skeletonema in Narra-
gansett Bay (USA): influence of food concentration and temperature. Marine
Biology 60, 101–113.
Desbonnet, A., Costa-Pierce, B. (Eds.), 2008. Science for Ecosystem-based Manage-
ment – Narragansett Bay in the 21st Century. Springer Verlag, New York, p. 570.
Doering, P., Oviatt, C.,1986. Application of filtration rate models to field populations
of bivalves: an assessment using experimental mesocosms. Marine Ecology
Progress Series 31, 265–275.
Doering, P., Oviatt, C., Kelly, J.,1986. The effects of the filter-feeding clam Mercenaria
mercenaria on carbon cycling in experimental mesocosms. Journal of Marine
Research 44, 839–861.
15N to Assess Nitrogen Sources and Fate in
Durbin, A., Durbin, E., 1981. Standing stock and estimated production rates of
phytoplankton and zooplankton in Narragansett Bay, Rhode Island. Estuaries 4,
Durbin, A., Durbin, E., 1988. Zooplankton and Ichthyoplankton in Narragansett Bay:
Status and Trends. Part 1: Zooplankton. Final Report to the New England
Interstate Water Pollution Control Commission for the Narragansett Bay
Project. Graduate School of Oceanography, University of Rhode Island, Narra-
Durbin, E., Krawiec, R., Smayda, T.,1975. Seasonal studies on the relative importance
of different size fractions of phytoplankton in Narragansett Bay (USA). Marine
Biology 32, 271–287.
Dwyer, R., 1980. Frequency Domain Studies of the Dynamics of an Estuarine
Ecosystem. Ph.D. thesis. Graduate School of Oceanography, University of Rhode
Island, Narragansett, RI, 158 pp.
Edwards, M., Richardson, A., 2004. Impact of climate change on marine phenology
and trophic mismatch. Nature 430, 881–884.
Eyre, B., Rysgaard, S., Dalsgaard, T., Christensen, P., 2002. Comparison of isotope
pairing and N2:Ar methods for measuring sediment denitrification – assump-
tions, modifications, implications. Estuaries 25, 1077–1087.
School of Oceanography, University of Rhode Island, Narragansett, RI, 62 pp.
Fish, C., 1925. Seasonal distribution of the plankton of the Woods Hole region.
Bulletin of the United States Bureau of Fisheries XLI, 91–175.
Frithsen, J., 1984. Metal incorporation by benthic fauna: relationships to sediment
inventory. Estuarine, Coastal and Shelf Science 19, 523–539.
Frithsen, J., 1989. The Benthic Communities Within Narragansett Bay. Final Report
to The Narragansett Bay Project, RI Dept. of Environmental Management,
Providence, RI. Graduate School of Oceanography, University of Rhode Island,
Narragansett, RI, 92 pp.
Fulweiler, R., 2007. The Impact of Climate Change on Benthic–Pelagic Coupling and
the Biogeochemical Cycling of Narragansett Bay, RI. Ph.D. thesis. Graduate
School of Oceanography, University of Rhode Island, Narragansett, RI, 250 pp.
Fulweiler, R., Nixon, S. Responses of benthic–pelagic coupling to climate change in
a temperate estuary. Hydrobiologia, in press.
Fulweiler, R., Nixon, S., Buckley, B., Granger, S., 2007. Reversal of the net dinitrogen
gas flux in coastal marine sediments. Nature 448, 180–182.
Fulweiler, R., Nixon, S., Buckley, B., Granger, S., 2008. Net sediment N2fluxes in
a coastal marine system – experimental manipulations and a conceptual model.
Furnas, M., Hitchcock, G., Smayda, T., 1976. Nutrient-phytoplankton relationships in
Narragansett Bay during the 1974 summer bloom. In: Wiley, M. (Ed.), Estuarine
Processes, vol. 1. Academic Press, NY, pp. 118–133.
Garber, J., 1982.15N Tracer and Other Laboratory Studies of Nitrogen Mineraliza-
tion in Sediments and Waters from Narragansett Bay, Rhode Island. Ph.D.
thesis. Graduate School of Oceanography, University of Rhode Island, Narra-
gansett, RI, 276 pp.
Gardner, W., McCarthy, M., An, S., Sobolev, D., Sell, K., Brock, D., 2006. Nitrogen
fixation and dissimilatory nitrate reduction to ammonium (DNRA) support
nitrogen dynamics in Texas estuaries. Limnology and Oceanography 51,
Gearing, J., Gearing, P., Rudnick, D., Requejo, A., Hutchins, M., 1984. The isotopic
variability of organic carbon in a phytoplankton-based temperate estuary.
Geochemica et Cosmochemica Acta 48, 1089–1098.
Gerritsen, J., Holland, A., Irvine, D.,1994. Suspension-feeding bivalves and the fate of
primary production: an estuarine model applied to Chesapeake Bay. Estuaries
Grassle, F., Grassle, J., Brown-Ledger, L., Petrecca, R., Copley, N., 1985. Subtidal
macrobenthos of Narragansett Bay: field and mesocosm studies of the effects of
eutrophication and organic input on benthic populations. In: Gray, J.,
Christiansen, M. (Eds.), Marine Biology of Polar Regions and Effects of Stress on
Marine Organisms. Wiley, New York, pp. 421–434.
Grebmeier, J., Overland, J., Moore, S., Farley, E., Carmack, E., Cooper, L., Frey, K.,
Helle, J., McLaughlin, F., McNutt, S., 2006. A major ecosystem shift in the
northern Bering Sea. Science 311, 1461–1464.
Greening, H., Janicli, A., 2006. Toward reversal of eutrophic conditions in
a subtropical estuary: water quality and seagrass response to nitrogen loading
reductions in Tampa Bay, Florida, USA. Environmental Management 38,
Greve, W., Prinage, S., Zidowitz, H., Nast, J., Reiners, F., 2005. On the phenology of
North Sea icthyoplankton. ICES Journal of Marine Science 62, 1216–1223.
Hale, S., 1974. The Role of Benthic Communities in the Nutrient Cycles of Narra-
gansett Bay. MS thesis. Graduate School of Oceanography, University of Rhode
Island, Narragansett, RI, 123 pp.
Hale, S., 1975. The role of benthic communities in the nitrogen and phosphorus
cycles of an estuary. In: Howell, F., Gentry, J., Smith, M. (Eds.), Mineral Cycling in
Southeastern Ecosystems. US Energy Research and Development Administra-
tion CONF-740513, pp. 291–308.
Henry, K., Nixon, S., 2008. A half century assessment of hard clam, Mercenaria
mercenaria, growth in Narragansett Bay, Rhode Island. Estuaries and Coasts, 31,
Hicks, S., 1959. The physical oceanography of Narragansett Bay. Limnology and
Oceanography 4, 316–327.
Hitchcock, G., Smayda, T.,1977. The importance of light in the initiation of the 1972–
1973 winter–spring bloom in Narragansett Bay. Limnology and Oceanography
S.W. Nixon et al. / Estuarine, Coastal and Shelf Science 82 (2009) 1–1816
Hopkinson, C., Smith, E., 2005. Estuarine respiration: an overview of benthic,
pelagic, and whole system respiration. Oxford. In: del Giorgio, P., Williams, P.
(Eds.), Respiration in Aquatic Ecosystems, pp. 122–146.
Howarth, R., Marino, R., Lane, J., Cole, J., 1988. Nitrogen fixation in freshwater,
estuarine, and marine ecosystems. Limnology and Oceanography 33, 669–687.
Husby, D., Nelson, C., 1982. Turbulence and vertical stability in the California
Current. CalCOFI Report 23, 113–129.
Iverson, R., 1990. Control of marine fish production. Limnology and Oceanography
Jeffries, H., 2001. Rhode Island’s ever changing Narragansett Bay. Maritimes 41, 1–5.
Jeffries, H., Terceiro, M., 1985. Cycle of changing abundances in the fishes of the
Narragansett Bay area. Marine Ecology Progress Series 25, 239–244.
Jeffries, P., Hale, S., Keller, A., 1989. Historical Data Assessment – Finfishes of the
Narragansett Bay Area. Final Report to the Narragansett Bay Project, RI Dept. of
Environmental Management, Providence, RI. Graduate School of Oceanography,
University of Rhode Island, Narragansett, RI, 407 pp.
Jossi, J., John, A., Sameoto, D., 2003. Continuous plankton recorder sampling off the
east coast of North America: history and status. Progress in Oceanography 58,
Kana, T., Darkangelo, C., Munt, M., Oldham, J., Bennett, E., Cornwell, J., 1994.
Membrane inlet mass-spectrometer for rapid high precision determination of
N2, O2, and Ar in environmental water samples. Analytical Chemistry 66,
Kana, T., Sullivan, M., Cornwell, J., Groskowski, K., 1998. Denitrification in estuarine
sediments determined by membrane inlet mass spectrometry. Limnology and
Oceanography 43, 334–339.
Karentz, D., Smayda, T., 1984. Temperature and the seasonal occurrence pattern of
30 dominant phytoplankton species in Narragansett bay over a 22-year period
(1959–1980. Marine Ecology Progress Series 18, 277–293.
Karentz, D., Smayda, T., 1998. Temporal patterns and variations in phytoplankton
community organization and abundance in Narragansett Bay during 1959–
1980. Journal of Plankton Research 20, 145–168.
Keller, A., 1988. Estimating phytoplankton productivity from light availability and
biomass in the MERL mesocosms and Narragansett Bay. Marine Ecology Prog-
ress Series 45, 159–168.
Keller, A., Riebesell, U., 1989. Phytoplankton carbon dynamics during a winter–
spring bloom in an enclosed marine ecosystem: primary productivity, biomass
and loss rates. Marine Biology 103, 131–142.
Keller, A., Frithsen, J., Oviatt, C., Maughan, J., Sullivan, B., Nixon, S., Pilson, M., 1987.
Marine Ecosystem Responses to Sewage Sludge and Inorganic Nutrient
Additions: a Mesocosm Experiment Data Report. MERL Series, Report No. 6.
Graduate School of Oceanography, University of Rhode Island, Narragansett,
RI, 237 pp.
Keller, A., Oviatt, C., Walker, H., Hawk, J., 1999. Predicted impacts of elevated
temperature on the magnitude of the winter–spring phytoplankton bloom in
temperate coastal waters: a mesocosm study. Limnology and Oceanography 44,
Kelly, J., 1983. Benthic–Pelagic Coupling in Narragansett Bay. Ph.D. thesis. Graduate
School of Oceanography, University of Rhode Island, Narragansett, RI, 195 pp.
Kelly, J., Nixon, S., 1984. Experimental studies of the effect of organic deposition on
the metabolism of a coastal marine bottom community. Marine Ecology Prog-
ress Series 17, 157–169.
Kelly, J., Berounsky, V., Nixon, S., Oviatt, C., 1985. Benthic-pelagic coupling and
nutrient cycling across an experimental eutrophication gradient. Marine
Ecology Progress Series 26, 207–219.
Kirby, R., Beaugrand, G., Lindley, J., Richardson, A., Edwards, M., Reid, P., 2007.
Climate effects and benthic–pelagic coupling in the North Sea. Marine Ecology
Progress Series 330, 31–38.
and biogeography of Skeletonema species (Bacillariophyta). Protist 159,177–193.
Kremer, J., Nixon, S., 1978. A Coastal Marine Ecosystem-Simulation and Analysis. In:
Ecological Studies 24. Springer Verlag, NY, 217 pp.
Levin, L., 1986. Effects of enrichment on reproduction in the opportunistic poly-
chaete Streblospio benedicti (Webster): a mesocosm study. Biological Bulletin
Li, Y., Smayda, T., 1998. Temporal variability of chlorophyll in Narragansett Bay,
1973–1990. ICES Journal of Marine Science 55, 661–667.
Libby, S., Borkman, D., Geyer, R., Keller, A., Mansfield, A., Turner, J., Anderson, D.,
Oviatt, C., Hyde, K., 2007. Water Column Monitoring in Massachusetts Bay:
1992–2006. Massachusetts Water Resources Authority Report ENQUAD 2007 –
11. Boston, 228 pp.
Loosanoff, V., 1939. Effect of temperature upon shell movements of clams, Venus
mercenaria (L). Biological Bulletin 76, 171–182.
Lucas, L., Cloern, J., Koseff, J., Monismith, S., Thompson, J., 1998. Does the Sverdurp
critical depth model explain bloom dynamics in estuaries? Journal of Marine
Research 56, 375–415.
MacIntyre, H., Geider, R., Miller, D., 1996. Microphytobenthos: the ecological role of
the ‘‘secret garden’’ of unvegitated shallow – water habitats. I Distribution,
abundance and primary production. Estuaries 19, 186–201.
Martin, J., 1966. Phytoplankton–Zooplankton Relationships in Narragansett Bay II:
the Seasonal Importance of Zooplankton Grazing and Nutrient Excretion. Ph.D.
thesis. Graduate School of Oceanography, University of Rhode Island, Narra-
gansett, RI, 99 pp.
Martin, J., 1970. Phytoplankton–zooplankton relationships in Narragansett Bay. IV.
The seasonal importance of grazing. Limnology and Oceanography 15, 412–418.
McCaffrey, R., Myers, A., Morrison, G., Bender, M., Ludtke, N., Cullen, D., Froelich, P.,
Klinkhammer, G.,1980. The relation between pore water chemistry and benthic
fluxes of nutrients and manganese in Narragansett Bay, Rhode Island.
Limnology and Oceanography 25, 31–44.
McCarthy, M., McNeal, K., Morse, J., Gardner, W., 2008. Bottom-water hypoxia
effects on sediment–water interface nitrogen transformations in a seasonally
hypoxic shallow bay (Corpus Christi May, TX, USA). Estuaries and Coasts 31,
Melrose, D., Oviatt, C., Berman, M., 2007. Hypoxic events in Narragansett Bay during
the summer of 2001. Estuaries and Coasts 30, 47–53.
Nay, J.,1996. Oligotrophication and its discontents: effects of reduced nutrient loading
on reservoir fisheries. American Fisheries Society Symposium 16, 285–295.
Niiler, P.P., Kraus, E., 1977. One-dimensional models of the upper ocean. In:
Krauss, E. (Ed.), Modelling and Prediction of the Upper Layers of the Ocean.
Pergamon Press, NY, pp. 1443–1472.
Nixon, S.,1981. Remineralization and nutrient cycling in coastal marine ecosystems.
In: Neilson, B., Cronin, L. (Eds.), Estuaries and Nutrients. Humana Press, Clifton,
NJ, pp. 111–138.
Nixon, S., 1988. Physical energy inputs and the ecology of lake and marine
ecosystems. Limnology and Oceanography 33, 1005–1025.
Nixon, S., 1989. An extraordinary red tide and fish kill in Narragansett Bay. In:
Cosper, E., Bricelj, E., Carpenter, E. (Eds.), Novel Phytoplankton Blooms. Springer,
NY, pp. 429–447.
Nixon, S., 1995. Coastal marine eutrophication: a definition, social causes, and
future concerns. Ophelia 41, 199–219.
Nixon, S., 1997. Prehistoric nutrient inputs and productivity in Narragansett Bay.
Estuaries 20, 253–261.
Nixon, S. Eutrophication and the macroscope. Hydrobiologia, in press.
Nixon, S., Pilson, M., 1984. Estuarine total system metabolism and organic exchange
calculated from nutrient ratios: an example from Narragansett Bay. In:
Kennedy, V. (Ed.), The Estuary as a Filter. Academic Press, NY, pp. 261–290.
Nixon, S., Oviatt, C., Hale, S., 1976. Nitrogen regeneration and the metabolism of
coastal marine bottom communities. In: Anderson, J., Macfadyen, A. (Eds.), The
Role of Terrestrial and Aquatic Organisms in Decomposition Processes. Black-
well, London, pp. 269–283.
Nixon, S., Oviatt, C., Kremer, J., Perez, K., 1979. The use of numerical models and
laboratory microcosms in estuarine ecosystem analysis- simulations of a winter
phytoplankton bloom. In: Dame, R. (Ed.), Marsh-Estuarine Systems Simulation.
University of South Carolina Press, Columbia, SC, pp. 165–188.
Nixon, S., Kelly, J., Furnas, B., Oviatt, C., Hale, S., 1980. Phosphorus regeneration and
the metabolism of coastal marine bottom communities. In: Tenore, K., Coull, B.
(Eds.), Marine Benthic Dynamics. University of South Carolina Press, Columbia,
SC, pp. 219–242.
Nixon, S., Pilson, M., Oviatt, C., Donaghay, P., Sullivan, B., Seitzinger, S., Rudnick, D.,
Frithsen, J., 1984. Eutrophication of a coastal marine ecosystem – an experi-
mental study using the MERL microcosms. In: Fasham, M. (Ed.), Flows of Energy
and Materials in Marine Ecosystems: Theory and Practice. Plenum Press, NY, pp.
Nixon, S., Granger, S., Nowicki, B.,1995. An assessment of the annual mass balance of
Nixon, S., Granger, S., Buckley, B., 2003. The warming of Narragansett Bay. Univer-
sity of Rhode Island. 41?N, The Magazine of the RI Sea Grant and Land Grant
Programs 2, 18–20. http://seagrant.gso.uri.edu/41N/vol2No1/baywarming.html.
Nixon, S., Granger, S., Buckley, B., Lamont, M., Rowell, B., 2004. A one hundred and
seventeen year coastal water temperature record from Woods Hole, Massa-
chusetts. Estuaries 27, 397–404.
Nixon, S., Buckley, B., Granger, S., Harris, L., Oczkowski, A., Cole, L., Fulweiler, R., 2005.
Anthropogenic Nutrient Inputs to Narragansett Bay – a Twenty-Five year perspec-
tive.Final Report to theNarragansett Bay Commissionand RISea Grant. Granduate
School of Oceanography, University of Rhode Island, Narragansett, RI, 50 pp.
Nixon, S., Buckley, B., Granger, S., Harris, L., Oczkowski, A., Fulweiler, R., Cole, L.,
2008. Nitrogen and phosphorus inputs to Narragansett Bay: past, present, and
future. In: Desbonnet, A., Costa-Pierce, B. (Eds.), Science for Ecosystem-based
Management, Narragansett Bay in the 21st Century. Springer, NY, pp. 101–175.
Nowicki, B., 1994. The effect of temperature, oxygen, salinity, and nutrient enrich-
ment on estuarine denitrification rates measured with a modified nitrogen gas
flux technique. Estuarine, Coastal and Shelf Science 38, 137–156.
Nowicki, B., Gold, A., 2008. Groundwater nitrogen transport and input along the
Narragansett bay coastal margin. In: Desbonnet, A., Costa-Pierce, B. (Eds.),
Science for Ecosystem-based Management: Narragansett Bay in the 21st
Century. Springer, NY, pp. 67–100.
Nowicki, B., Nixon, S., 1985. Benthic community metabolism in a coastal lagoon
ecosystem. Marine Ecology Progress Series 22, 21–30.
Oczkowski, A., Nixon, S., Henry, K., DiMilla, P., Pilson, M., Granger, S., Buckley, B.,
Thornber, C., McKinney, R., Chaves, J., 2008. Distribution and trophic impor-
tance of anthropogenic nitrogen in Narragansett Bay: an assessment using
stable isotopes. Estuaries and Coasts 31, 53–69.
Olsen, S., Robadue, D., Lee, V., 1980. An Interpretive Atlas of Narragansett Bay. In:
University of Rhode Island Marine Bulletin 40. Coastal Resources Center, 82 pp.
O’Reilly, J., 1984. Carbon Flow in a Coastal Marine Bottom Community. Ph.D.
thesis. Graduate School of Oceanography, University of Rhode Island, Narra-
gansett, RI, 199 pp.
Oviatt, C., 1981. Effectsof different
zooplankton and nutrients in marine microcosms. Marine Ecology Progress
Series 4, 57–67.
mixingschedules on phytoplankton,
S.W. Nixon et al. / Estuarine, Coastal and Shelf Science 82 (2009) 1–1817
Oviatt, C., 2004. The changing ecology of temperate waters during a warming trend. Download full-text
Estuaries 27, 895–904.
Oviatt, C., Nixon, S., 1975. Sediment resuspension and deposition in Narragansett
Bay. Estuarine and Coastal Marine Science 3, 201–217.
Oviatt, C., Buckley, B., Nixon, S., 1981. Annual phytoplankton metabolism in
Narragansett Bay calculated from survey field measurements and microcosm
observations. Estuaries 4, 167–175.
Oviatt, C., Pilson, M., Nixon, S., Frithsen, J., Rudnick, D., Kelly, J., Grassle, J., Grassle, J.,
1984. Recovery of a polluted estuarine ecosystem: a mesocosm experiment.
Marine Ecology Progress Series 16, 203–217.
Oviatt, C., Doering, P., Nowicki, B., Reed, L., Cole, J., Frithsen, J., 1995. An ecosystem
level experiment on nutrient limitation in temperate coastal marine environ-
ments. Marine Ecology Progress Series 116, 171–179.
Oviatt, C., Keller, A., Reed, L., 2002. Annual primary production in Narragansett Bay
with no bay-wide winter–spring bloom. Estuarine, Coastal and Shelf Science 54,
Oviatt, C., Olsen, S., Andrews, M., Collie, J., Lynch, T., Raposa, K., 2003. A century of
fishing and fish fluctuations in Narragansett Bay. Reviews in Fisheries Science
Oviatt, C., Hyde, K., Keller, A., Turner, J., 2007. Production patterns in Massachusetts
Bay with outfall relocation. Estuaries and Coasts 30, 35–46.
Ozaki, K., Uye, S., Kusumoto, T., Hagino, T., 2004. Interannual variability of the
ecosystem of the Kii Channel, the Inland Sea of Japan, as influenced by bottom
intrusion of cold and nutrient-rich water from the Pacific Ocean, and a recent
trend of warming and oligotrophication. Fisheries Oceanography 13, 65–79.
Parsons, T., Takahashi, M., 1973. Biological Oceanographic Processes. Pergamon
Petersen, J., 2004. Grazing on pelagic primary producers-the role of benthic
suspension feeders in estuaries. In: Nielsen, S., Banta, G., Pedersen, M. (Eds.),
Estuarine Nutrient Cycling: the Influence of Primary Producers. Kluwer, Boston,
MA, pp. 129–152.
Phelps, D.,1958. A Quantitative Study of the Infauna of Narragansett Bay in Relation
to Certain Physical and Chemical Aspects of their Environments. MS thesis.
Graduate School of Oceanography, University of Rhode Island, Narragansett, RI,
Philippart, C., Beukema, J., Cadee, G., Dekker, R., Goedhart, P., vanIperen, J.,
Leopold, M., Herman, P., 2007. Impacts of nutrient reduction on coastal
communities. Ecosystems 10, 95–118.
Pilling, G., Miller, R., Easey, M., Maxwell, D., Tidd, A., 2007. Phenology and North Sea
cod Gadus morhua L.: has climate change affected otolith annulus formation
and growth? Journal of Fish Biology 70, 584–599.
Pilson, M., 1985. On the residence time of water in Narragansett Bay. Estuaries 8,
Pilson, M., 2008. Narragansett Bay amidst a globally changing climate. In: Des-
bonnet, A., Costa-Pierce, B. (Eds.), Science for Ecosystem-based Management:
Narragansett Bay in the 21st Century. Springer, NY, pp. 35–46.
Pilson, M., Oviatt, C., Nixon, S., 1980. Annual nutrient cycles in a marine microcosm.
In: Giesy, J. (Ed.), Microcosms in Ecological Research. US Department of Energy,
Symposium Series 52 (CONF-781101), pp. 753–778.
Pratt, D., 1959. The phytoplankton of Narragansett Bay. Limnology and Oceanog-
raphy 4, 425–440.
Pratt, D., 1965. The winter–spring diatom flowering in Narragansett Bay. Limnology
and Oceanography 10, 173–184.
Pryor, D., Saarman, E., Murray, D., Prell, W., 2007. Nitrogen Loading from Waste-
water Treatment Plants to Upper Narragansett Bay. Narragansett bay Estuary
Program Report NBEP-2007-122, 36 pp.
Rabalais, N., Nixon, S., 2002. Nutrient over-enrichment in coastal waters: global
patterns of cause and effect. Estuaries 25, 4B.
Richardson, A., Schoeman, D., 2004. Climate impact on plankton ecosystems in the
Northeast Atlantic. Science 305, 1609–1612.
Riebesell, U.,1989. Comparison of sinking and sedimentation rate measurements in
a diatom winter/spring bloom. Marine Ecology Progress Series 54, 109–119.
Riisgård, H., 2004. Zoobenthic consumers of the primary production in the marine
environment. Recent Research Developments in Environmental Biology 1,
Riley, G., 1967. The plankton of estuaries. In: Lauff, G. (Ed.), Estuaries. American
Association for the Advancement of Science Publication, No. 83, pp. 316–328.
Rogers, J., 2008. Circulation and transport in upper Narragansett Bay. MS thesis.
Graduate School of Oceanography, University of Rhode Island, Kingston, RI,
Rudnick, D., Oviatt, C., 1986. Seasonal lags between organic carbon deposition and
mineralization in marine sediments. Journal of Marine Research 44, 815–837.
Rudnick, D., Elmgren, R., Frithsen, J., 1985. Meiofaunal prominence and benthic
seasonality in a coastal marine ecosystem. Oecologia 67, 157–168.
Sarno, D., Wiebe, Kooistra, W., Medlin, L., Percopo, I., Zingone, A., 2005. Diversity in
the genus Skeletonema (Bacillariophyceae). II. An assessment of the taxonomy of
S. costatum-like species with the description of four new species. Journal of
Phycology 41, 151–176.
Schell, D., 2000. Declining carrying capacity in the Bering Sea: isotopic evidence
from whale baleen. Limnology and Oceanography 45, 459–462.
Seitzinger, S.,1982. The Importance of Denitrification and Nitrous Oxide Production
in the Nitrogen Dynamics and Ecology of Narragansett Bay, Rhode Island. Ph.D.
thesis. Graduate School of Oceanography, University of Rhode Island, Narra-
gansett, RI, 145 pp.
Seitzinger, S., Garber, J.,1987. Nitrogen fixation and15N2calibration of the acetylene
reduction assay in coastal marine sediments. Marine Ecology Progress Series 37,
Seitzinger, S., Giblin, A., 1996. Estimating denitrification in North Atlantic conti-
nental shelf sediments. Biogeochemistry 35, 235–260.
Seitzinger, S., Nixon, S., Pilson, M., Burke, S., 1980. Denitrification and N2O
production in near-shore marine sediments. Geochimica et Cosmochimica Acta
Seitzinger, S., Nixon, S., Pilson, P.,1984. Denitrification and nitrous oxide production
in a coastal marine ecosystem. Limnology and Oceanography 29, 73–83.
Sisson, R., 1973. Biological and Commercial Fisheries Related Research on the
Channeled Welk Busycon canaliculatum (Linne) in Narragansett Bay, Rhode
Island. MS thesis. University of Rhode Island, Kingston, RI, 62 pp.
Smayda, T., 1955. Phytoplankton Studies in Lower Narragansett Bay. MS thesis.
Graduate School of Oceanography, University of Rhode Island, Narragansett,
RI, 114 pp.
Smayda, T., 1957. Phytoplankton studies in lower Narragansett Bay. Limnology and
Oceanography 11, 342–359.
Smayda, T.,1973. The growth of Skeletonema costatum during a winter–spring bloom
in Narragansett Bay, Rhode Island. Norwegian Journal of Botany 20, 219–247.
Smayda, T., 1998. Patterns of variability characterizing marine phytoplankton, with
examples from Narragansett Bay. ICES Journal of Marine Science 55, 562–573.
Smayda, T., Borkman, D., 2008. Nutrient and plankton dynamics in Narragansett
Bay. In: Desbonnet, A., Costa-Pierce, B. (Eds.), Science for Ecosystem-based
Management, Narragansett Bay in the 21st Century. Springer, NY, pp. 431–484.
Smetacek, V.,1984. The supply of food to the benthos. In: Fasham, M. (Ed.), Flows of
Energy and Materials in Marine Ecosystems. NATO Conference Series. Plenum
Press, NY, pp. 517–547.
Smetacek, V., Pollehne, F.,1986. Nutrient cycling in pelagic systems: a reappraisal of
the conceptual framework. Ophelia 26, 401–428.
Stickney, A., Stringer, L., 1957. A study of the invertebrate bottom fauna of Green-
wich Bay, R.I. Ecology 38, 111–122.
Steemann-Nielsen, E., Hansen, V., 1959. Light adaptation in marine phytoplankton
populations and its interrelation with temperature. Physiologia Plantarum 12,
Stenseth, N., Ottersen, G., Hurrell, J., Belgrano, A., 2004. Marine Ecosystems and
Climate Variation – the North Atlantic Perspective. Oxford University Press.
Suffolk County Department of Health Services, 2007. Response to Public Comments
on Total Maximum Daily Load for Nitrogen in the Peconic Estuary, 27 pp.
Available from:. Prepared in Cooperation with the U.S. EPA, NY State Depart-
ment of Environmental Conservation and the Peconic Estuary Program http://
Sullivan, B., Van Keuren, D., Clancy, M., 2001. Timing and size of blooms of the
ctenophore Mnemiopsis leidyi in relation to temperature in Narragansett Bay,
Rhode Island. Hydrobiologia 451, 113–120.
Taylor, D., 2003. Size-dependent predation on post-settlement winter flounder,
Pseudopleuronectes americanus by sand shrimp, Crangon septemspinosa. Marine
Ecology Progress Series 263, 197–215.
Townsend, D., Cammen, L., 1988. Potential importance of the timing of spring
plankton blooms to benthic–pelagic coupling and recruitment of juvenile
demersal fishes. Biological Oceanography 5, 215–229.
Tucker, J., Kelsey, S., Giblin, A., Hopkinson, C., 2008. 2007 Annual Benthic Nutrient
Flux Monitoring Report. Massachusetts Water Resources Authority. Report
ENQUAD 2008–12. Boston, 59 pp.
Turner, J., Borkman, D., Lincoln, J., Gauthier, D. and Petitpas, C., Plankton studies in
Buzzards Bay, Massachusetts, USA. VI. Phytoplankton and water quality,1987 to
1998. Marine Ecology Progress Series, in press.
Valiela, I., 1995. Marine Ecological Processes, second ed. Springer, NY, 686 pp.
Valiela, I., 2006. Global Coastal Change. Blackwell Publishing, Malden, MA.
Vargo, G., 1976. The Influence of Grazing and Nutrient Excretion by Zooplankton on
the Growth and Production of the Marine Diatom, Skeletonema costatum
(Greville) Cleve, in Narragansett Bay. Ph.D. thesis. Oceanography, University of
Rhode Island, Narragansett, RI, 118 pp.
Vargo, G., 1979. The contribution of ammonia excreted by zooplankton to phyto-
plankton production in Narragansett Bay. Journal of Plankton Research 1, 75–84.
Vinogradov, A., 1953. The Elementary Composition of Marine Organisms (J. Efron, J.
Setlow, Trans.). Sears Foundation for Marine Research, Yale University, New
Haven, CT, 647 pp.
Ware, D., Thomson, R., 21 April 2005. Bottom-up ecosystem trophic dynamics
determine fish production in the northeast Pacific. Sciencexpress, 10. 10.1126/
Wlodarczyk, E., Durbin, A., Durbin, E., 1992. Effect of temperature on lower feeding
thresholds, gut evacuation rate, and diel feeding behavior in the copepod
Acartia hudsonica. Marine Ecology Progress Series 85, 93–106.
Yoder, J., 1979. Modeling the Effect of Light Intensity and Temperature on the
Growth Rate and Biomass of Skeletonema costatum (Bacillariophyceae) and
other Diatoms. Ph.D., Graduate School of Oceanography, University of Rhode
Island, Narragansett, RI, 145 pp.
York, J., Quay, P., Valiela, I., 2007. Stable isotopic detection of ammonium and nitrate
assimilation by phytoplankton in the Waquoit Bay estuarine system. Limnology
and Oceanography 52, 144–155.
S.W. Nixon et al. / Estuarine, Coastal and Shelf Science 82 (2009) 1–1818