ti
l
an
er, P
04;
line
ic e
in water clarity. Time series of absorption and scattering coe?cients were measured at 1-h intervals for nearly 2 years. The seasonal
ncepattern in weekly averaged absorption and scattering coe?cients in each year was driven primarily by changes in the particulate
matter of both biogenic and mineral origin. Temporal patterns in particulate absorption and scattering resulted from identi?able
events that di?ered in relative magnitude between the 2 years: a spring bloom was followed by a transient ??clear-water?? phase,
followed by increases in non-algal particulate matter to a late-summer maximum, and rapid declines of all parameters in late fall.
Interannual variability in the spring bloom was governed by timing and magnitude of nutrient inputs from the watershed, while
major patterns in summer variability of both organic and inorganic particulate matter appeared to follow the general cycle of
biological activity in the system.

 2005 Elsevier Ltd. All rights reserved.
Keywords: optical properties; estuary; absorption coe?cient; scattering coe?cient; suspended solids
1. Introduction
The penetration of light underwater is of fundamen-
tal importance to aquatic ecosystems. The quantity and
quality of underwater light drive photosynthesis by
phytoplankton, benthic algae, and submersed macro-
phytes (Kirk, 1994). Light is required for feeding by
visually orienting predators (Aksnes et al., 2004), and
the visual appearance of a water body is a key factor
in the aesthetic and recreational appeal to humans
(Davies-Colley et al., 1993). Understanding the factors
controlling the variability of optical properties in coastal
waters is important, because changes in the optical
properties of a water body can be both indicators of
stressors to components of the ecosystem, e.g. sub-
merged aquatic vegetation (SAV), that require high light
intensities for survival.
In coastal watersheds worldwide, human population
growth has been increasing at a high rate and has
negatively impacted living resources at many locations
(Richardson and J?rgensen, 1996; Rabalais and Turner,
2001). Many of the processes that threaten estuarine
habitat quality result in detectable changes to the optical
properties of the water. For example, eutrophication
results in increasing frequency and magnitude of
phytoplankton blooms that absorb and scatter light in
spectrally identi?able ways (Gallegos and Jordan, 2002).
Highway construction and certain agricultural practicesTemporal variability of op
eutrophic estuary: Seasona
C.L. Gallegos*, T.E. Jord
Smithsonian Environmental Research Cent
Received 20 September 20
Available on
Abstract
We monitored inherent optical properties in a turbid, eutroph
Estuarine, Coastal and Shelf Sciestressors such as eutrophication, and can be potential
* Corresponding author.
E-mail address: gallegosc@si.edu (C.L. Gallegos).
0272-7714/$ - see front matter 
 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.ecss.2005.01.013cal properties in a shallow,
and interannual variability
, A.H. Hines, D.E. Weller
.O. Box 28, Edgewater, MD 21037, USA
accepted 28 January 2005
2 April 2005
stuary to determine the factors a?ecting the temporal variability
64 (2005) 156e170
www.elsevier.com/locate/ECSSincrease erosion and delivery of suspended sediments
that increase the scattering and re?ectivity of estuarine
waters (Stumpf and Pennock 1989, 1991). Altered ?ow
paths from wetlands and certain industrial discharges
(e.g. paper mill e?uents) increase the concentration of
C.L. Gallegos et al. / Estuarine, Coastalcolored dissolved organic matter (CDOM) of some
estuaries, greatly increasing the absorption of blue and
ultraviolet light (McPherson and Miller 1987; Gallegos
and Kenworthy, 1996). Thus changes in the optical
properties of estuarine waters are a potentially sensitive
indicator of stressors characteristic of human distur-
bance. Long series of observation, however, are needed
to discern human induced changes from normal seasonal
and interannual variability.
Estuaries are optically complex systems due to the
sources, sinks, and transformations of terrestrially
derived dissolved and particulate matter that mixes with
coastal-derived water that is likewise optically complex
(Davies-Colley et al., 1993). Flocculation of river-borne
clay particles occurs with small increases in salt con-
centration, changing the e?ective size distribution and
settling rate of suspended solids (Postma, 1967).
Interactions between particulate matter and biological
production also occur on longer (i.e. seasonal) time
scales, resulting in, e.g. co-?occulation of phytoplankton
and inorganic mineral particles (Stross and Sokol, 1989).
A variety of physical (Sanford, 1994) and biological
(Hines et al., 1990) processes operate to resuspend
bottom sediments on a variety of time scales. Such
processes change not only the concentration of optically
active particulate matter in the water column, but also
alter the size-distributions of particulate matter, thereby
a?ecting the absorption and scattering per unit mass
of suspended solids (Mobley, 1994). Sorting out the
relative timing and magnitude of such in?uences on
estuarine optical properties is challenging due to the
complexity of processes and di?culty in making the
relevant measurements.
Advances in instrumentation have made it possible to
monitor changes in inherent optical properties of water
bodies (Dickey et al., 1998; Chang and Dickey, 2001),
i.e. those properties that depend only on the amounts
and kinds of material in the water. Absorption and
scattering spectra are especially informative because
from them it is possible to estimate concentrations of
materials in the water responsible for absorption and
attenuation (Chang and Dickey, 2001; Gallegos and
Neale, 2002). For example, Chang and Dickey (1999)
examined time series of absorption and scattering
spectra from moored instrumentation o? Georges Bank,
and observed high concentrations of suspended solids in
a near-bottom mooring following the passage of
a hurricane. Gallegos and Jordan (2002) measured
changes in absorption and scattering spectra during
the course of a large bloom of the dino?agellate
Prorocentrum minimum. They observed a three-fold
increase in attenuation due to the bloom and determined
that the optical e?ects of the bloom were initially
dominated by phytoplankton absorption, and then by
non-algal detrital matter at the termination of the
bloom.Here we present the results of a nearly 2-year series of
hourly measurements of inherent optical properties in
the Rhode River. Our objectives were to determine the
magnitude and variability of factors responsible for
reduced water clarity in this systemwhich once supported
an abundance of submerged vegetation (Southwick
and Pine, 1975). These observations are unique in their
coverage of the seasonal and interannual variability, and
in their resolution of the components of absorption. Key
features of the variability are interannual di?erences in
the magnitude of the spring bloom, and the relative
constancy of the late summer peak of non-algal par-
ticulate matter. Additionally, we examine a variety of
long-term ancillary data to propose causal mechanisms
for the seasonal pattern of non-algal particulates.
2. Study site
Inherent optical properties were monitored at a site on
the Rhode River (38.883 
N, 76.533 
W), Maryland. The
Rhode River is a small, shallow tributary embayment, or
subestuary, on the western shore of Chesapeake Bay
(Fig. 1). The subestuary is located in the mesohaline zone
of Chesapeake Bay, where the salinity varies seasonally
from 5 to 18 at the mouth depending on ?ow of the
Susquehanna River, and from 0 to 14 at the head,
depending on ?ow of the Susquehanna River and local
?ow of Muddy Creek, the principal freshwater source to
the Rhode River. Depth varies from about 4 m at the
mouth to !1 m at an area of subtidal mud ?ats up-
estuary of the monitoring station. The subestuary is
usually vertically homogenous, though subsurface chlo-
rophyll accumulations sometimes occur during dense
phytoplankton blooms (Gallegos et al., 1990).
The subestuary is eutrophic, with summer chlorophyll
concentrations averaging 20e40 mgm3, and experien-
ces extraordinary spring blooms of the dino?agellate
Prorocentrum minimum in years when the spring freshet
of the Susquehanna River is su?cient to infuse nitrate
into the system in mid- to late April (Gallegos et al.,
1990, 1997; Gallegos and Jordan, 2002). Inputs of
suspended solids from the local watershed peak in spring
(Jordan et al., 1986), though concentrations of sus-
pended solids in the water column peak in late summer
(Jordan et al., 1991, and see below).
3. Materials and methods
3.1. Optical monitoring
Spectral absorption [a(l)] and attenuation [c(l)]
coe?cients were measured at 9 wavelengths: 412, 440,
488, 510, 532, 555, 650, 676, and 715 nm, using a ?ow-
157and Shelf Science 64 (2005) 156e170through absorption-attenuation meter (ac9, Wetlabs).
lFlow to the ac9 was by gravity feed. Water from the
estuary was pumped into an open topped polyvinyl
chloride (PVC) cylinder with an over?ow pipe to
provide a constant head. A spigot and tubing located
near the bottom of the cylinder conducted water past
a bubble trap standpipe to the ac9. A Wetlabs MPak
was used to control the ac9 and log the data. Once per
hour the ac9 was turned on, allowed to warm up for
5 min, and sampled for 1 min at 6 Hz.
The PVC cylinder, ac9, and MPak were housed in
a monitoring shed at the end of the Smithsonian pier
on the Rhode River (Fig. 1). To reduce fouling within
the cylinder, the pump was operated only 15 min h1,
extending from about 10 min before to 5 min after
sampling by the ac9. At the conclusion of a sampling
cycle, a solenoid valve was opened that allowed about
60 ml of bromide solution to ?ow through the standpipe
and ac9 to inhibit growth of fouling organisms on the
optical surfaces of the ac9. Timing of the pump and
solenoid valve was controlled by a Campbell Scienti?c
CR10 data logger and control module.
The system was not operated when the estuary was
January and February of 2000 when ice cover was
prolonged, we augmented the series by collecting grab
samples through a hole in the ice for 23 days. These
samples were run through an ac9 in the laboratory as
described by Gallegos and Neale (2002).
3.2. Data processing
Datawere downloaded and the ac9was cleaned 3 times
per week. Cleaning was commenced immediately follow-
ing a reading. Water from that reading was collected and
run through the meter again after cleaning to correct for
drift due to fouling or accumulation of particulate matter
during the previous monitoring period. Cleaning and
downloading generally took less than 1 h, so that hourly
sampling was usually not interrupted.
We calculated the change in absorption [a(l)] and
beam attenuation [c(l)] coe?cients after cleaning and
applied the di?erence incrementally over the time period
since the last cleaning, which was generally !3 d. We
calculated the increment linearly from late fall to early
spring when drift was due to settling of inanimate
Fig. 1. Map of the Rhode River, Maryland, showing locations of optical monitoring station, bottle-cast stations, water quality transect, and trawl
lines. Inset shows location of the Rhode River on the western shore of Chesapeake Bay.158 C.L. Gallegos et al. / Estuarine, Coastaice-covered, to avoid damage to the sampling pump. Inand Shelf Science 64 (2005) 156e170particulate matter in the tubes, and exponentially during
al aspring and summer when drift was due primarily to
biological fouling. Inappropriate application of a linear
correction could be recognized by the calculation of
negative coe?cients near the beginning of the monitor-
ing period. Temperature and salinity at the site were
monitored at hourly intervals coinciding with ac9
measurements by a Hydrolab () Minisonde. These
measurements were used to calculate corrections for the
temperature and salinity dependence of the absorption
coe?cient of pure water according to the manufacturer?s
speci?cations. In these waters, the resulting corrections
were always small in relation to the overall magnitude
and relative variability in measured optical coe?cients.
3.3. Data analysis
We estimated the components of absorption by a
modi?cation of the procedure of Gallegos and Neale
(2002). Brie?y, the absorption coe?cient referenced to
water, at  w(l), at wavelength lmay be represented as the
sum of absorption due to colored dissolved organic
matter (CDOM), phytoplanktonpigments, andnon-algal
particulate matter,
atw?l?Zag?l?Caf?l?Capf?l? ?1?
where the subscripts refer to absorption by particular
components: g for CDOM (i.e. gelbsto?, Kirk, 1994), f
for phytoplankton, and p  f for non-algal particulate
matter. The average spectral shape of absorption by a
component was represented by a normalized absorption
function, determined by dividing the absorption by a
component at each wavelength by the value at
a reference wavelength. The normalized absorption
spectra were determined using laboratory measurements
in which the components were physically separated, as
reported previously (Gallegos and Neale, 2002). The
estimation procedure determines the value of absorption
by each component at the reference wavelength, from
which absorption at all other wavelengths may be
calculated by multiplication by the normalized absorp-
tion function. Reference wavelengths were chosen as
440 nm for CDOM and non-algal particulates by
convention (Kirk, 1994), and 676 nm for phytoplankton
because it is at an absorption peak of chlorophyll.
Further details and the modi?cations to the procedure
of Gallegos and Neale (2002) are given in Appendix A.
We represented the wavelength dependence of partic-
ulate scattering spectrum as the product of an exponen-
tial function of wavelength scaled by the particulate
scattering at a reference wavelength (555 nm),
bp?l?Zbp?555?


555
l
h
?2?
where bp(l)Z the particulate scattering spectrum, and h
C.L. Gallegos et al. / Estuarine, Coastis the scattering spectral exponent. We estimated h asthe negative slope of a logelog regression of the 9
measured bp(l) against l. Eq. (2) well characterizes the
spectrum when scattering is dominated by particles
having sizes smaller or comparable to the wavelength of
light, whereas when scattering is dominated by larger
particles, estimated h is small and wavelength de-
pendence of scattering is characterized by depressions
associated with the red and blue absorption peaks of
chlorophyll (Stramski and Mobley, 1997). Scattering by
pure water and by CDOM are considered negligible
(Kirk, 1994).
3.4. Laboratory measurements of optical properties
To develop the normalized absorption functions of
the optically active components and to examine spatial
gradients in optical properties, we collected vertically
integrated whole-water samples from 3 stations: up-
estuary of the optical monitoring station in a region of
subtidal mud?ats, near the monitor, and down estuary
of the monitor at the mouth of the Rhode River (Fig. 1,
open squares). Components of absorption and the
normalized absorption spectra were measured as de-
scribed previously (Gallegos and Neale, 2002). Here we
present measurements of absorption by CDOM at
440 nm [i.e. ag(440)] as an independent check on values
estimated by inversion of monitored ac9 data (Gallegos
and Neale, 2002, see Appendix A), and for delineating
spatial gradients as an indication of CDOM sources.
3.5. Supporting measurements
To gain insight into the processes governing seasonal
variations in optical properties, we compared patterns of
variation in optical properties with various long-term
measurements of parameters capable of in?uencing
CDOM, phytoplankton, and total suspended solids
(TSS). We measured the concentration of TSS in a series
of transects of the Rhode River and Muddy Creek at
approximately bi-weekly intervals from 1986 to 2001.
The concentration of TSS was determined from the
weight gain of a pre-weighed, 0.45 mm Poretics ?lter.
Additionally, during 2000 and 2001 we determined the
fraction of combustible suspended solids by ?ltering
samples through a pre-combusted GF/F glass ?ber ?lter,
and determining the dry-weight loss after combustion
for 3 h at 500 
C. Methods of sample collection and
data through 1989 are reported in Jordan et al. (1991).
Concentrations of NO3, NH4, and PO4 in samples from
the bi-weekly transects were measured by previously
published methods (Jordan et al., 1991). Data shown
here were from the transect segment closest the optical
monitoring station (Fig. 1, dotted line).
Discharge from 36% of the watershed of Muddy
Creek was monitored from 1985 to 1999 with a network
159nd Shelf Science 64 (2005) 156e170of weirs and automated samplers. Location of the
l aautomated samplers is given in Gallegos et al. (1992).
Flow was measured at V-notch weirs, and samples in
volumes proportional to the ?ow were pumped into
acid-washed glass carboys (Correll, 1981; Correll et al.,
1999). Samples were composited weekly and analyzed
for TSS concentration as described above. These data
provided an estimate of the seasonal pattern of weekly
mean suspended sediment in?ux from the local water-
shed (Jordan et al., 1986).
We estimated the contribution of living plankton to
the organic fraction of TSS from long-term measure-
ments of phytoplankton chlorophyll and ciliate abun-
dance. Vertically integrated water samples were collected
from a station ca. 100 m from the monitoring pier using
a 2-l Labline Te?on sampler by slowly lowering and
raising the sampler in less time than required to ?ll the
sampler. Samples for chlorophyll analysis (1990e2001)
were ?ltered onto Whatman GF/F glass ?ber ?lters,
extracted in 10 ml of 90% acetone overnight at 4 
C
either immediately or after freezing for !2 week. We
estimated the dry weight of phytoplankton by assuming
a carbon-to-chlorophyll ratio of 40 mg C (mg Chl)1
(Gallegos, 2001) and a dry weight-to-carbon ratio of
3.33 mg dry weight (mg C)1 based on Red?eld
elemental composition. Samples for ciliate counting
(1992e1994) were preserved in modi?ed Bouin?s ?xative
(Coats and Heinbokel, 1982) and processed by the
quantitative protargol staining (QPS) technique of
Montagnes and Lynn (1993). Ciliate abundances were
derived from 25e50 ml QPS preparations by enumerat-
ing specimens present in arbitrarily selected microscope
?elds (!1250) until a total of 100 individuals or the
equivalent of 2 ml of whole-water sample was counted.
Ciliates were classi?ed by morphotypes, with genus and
species identi?cations made whenever possible. Length
and width measurements were obtained for each
morphotype using a calibrated ocular micrometer. For
common ciliate taxa, ?ve length-width determinations
were made per sample until R100 measurements were
obtained. Uncommon taxa were sized when encountered
(10e50 estimates per morphotype); however, !10
measurements were obtained for rare morphotypes.
Species-speci?c biovolumes were calculated from QPS
dimensions using appropriate geometric formulae.
For each sample, species abundance and biovolume
were multiplied, and resultant values summed across
morphotypes to give cumulative biovolumes for total
ciliates. Cumulative biovolumes were converted to
carbon biomass using a factor of 0.27 pg C mm3 for
protargol stained specimens (Bockstahler and Coats,
1993), and to dry weight by a factor of 2.11 g dry weight
(g C)1 based on Fenchel and Finlay (1983).
As an indicator of the potential for bioturbation to
in?uence TSS concentrations, we examined the monthly
averaged biomass of epibenthic ?sh and crabs in the
160 C.L. Gallegos et al. / Estuarine, CoastaRhode River. Abundances of epibenthic ?sh and crabswere estimated with otter trawls from 1981 to 2002 by
procedures of Hines et al. (1990). For each trawl (Fig. 1,
arrows), all organisms caught were counted and identi-
?ed to species. Body sizes (total body length of ?sh and
maximal carapace width of blue crabs) of a subset of 20
individuals were randomly selected for measurement;
all individuals were measured when total catch of
a species was !20 individuals. Biomass as wet weight
was calculated from established regressions against
length (Dawson, 1965 for ?sh; Olmi and Bishop, 1983
for crabs), and expressed as kg trawl1. Analysis was
restricted to the most consistently observed mobile
bottom feeding taxa, namely blue crabs (Callinectes
sapidus), and demersal ?sh mainly composed of spot
(Leiostomus xanthurus), and croaker (Micropogonias
undulatus).
4. Results
4.1. Monitoring e?ciency
Data from the monitoring system consist of 10,534
measurements of spectral absorption and attenuation
coe?cients measured from January 2000 through Sep-
tember 2001. Excluding days that the system was not
operated due to the threat of ice (91 days between the two
winters), gross e?ciency of data collection was approx-
imately 80%. Data loss occurred due to a combination of
occasional recorder failure, loss of power, and 431
measured spectra were rejected due to entrainment of
bubbles or accumulation of particulate matter, which
were recognized by high and erratic spectra that pro-
duced unrealistic estimates of absorption components.
The data show a high degree of variability on nearly
every time scale. Therefore we calculated weekly averages
for better resolution of seasonal and interannual
patterns. Shorter-term variability will be examined in
another paper.
4.2. Absorption spectra
Weekly averages of total absorption coe?cient at
440 nm and the magnitudes of the components com-
prising it underwent a similar pattern of variation in
each year (Fig. 2a and Fig. 3a). There was a minimum
for all components in late winter, a noticeable spring
phytoplankton bloom (ca. day 120), a decline in all
components following the collapse of the spring bloom
(ca. day 150), a gradual build up of non-algal particulate
absorption for the remainder of the summer, and a rapid
decline of all components in late summer, ca. day 270
(only measured in 2000).
Absorption by CDOM was the least variable absorp-
tion component, varying from about 0.3 m1 in winter to
1
nd Shelf Science 64 (2005) 156e170about 1 m in summer (Fig. 2a). There was a tendency
al afor the inversion algorithm to overestimate ag(440),
though the overall seasonal pattern was well resolved in
2000. Due to the inability to predict the absorption-to-
scattering ratio (r) exactly (see Appendix A), there is an
unavoidable tendency for estimates of ag(440) to covary
with estimated ap  f(440) (Fig. 2a and Fig. 3a, see also
Gallegos and Neale, 2002). This covariation between
estimates of ag(440) and ap  f(440) resulted in, for
example, overestimates of ag(440) following the spring
phytoplankton bloom in 2000 (Fig. 2a, days 135e150)
and underestimate ag(440) after a fall phytoplankton
bloom in 2001 (Fig. 3a, days 195e210).
The spring phytoplankton bloom was much larger in
2000 than in 2001, but in each year there was an increase
150 180 210 240 270 300 330 3600 30 60 90 120
150 180 210 240 270 300 330 3600 30 60 90 120
0
1
2
3
4
5
6
7
8
a
p-
(440)
0
2
4
6
8
10
12
14
16
18
20
22
Day of Year 2000
-1.0
-0.8
-0.6
-0.4
-0.2
0.0
0.2
0.4
0.6
0.8
1.0
a
b
A
bs
or
pt
io
n 
C
oe
ffi
ci
en
t (
m
-1
)
J F M A M J J A S O N D
J F M A M J J A S O N D
b
p
(5
55
) 
(m
-1
) 
a
g
(440) 
 Obs. a
g
(440)
a
  
(440)
b
p
(555)
Fig. 2. (a) Weekly averages of components of absorption at 440 nm in
the Rhode River estimated from monitored absorption and attenua-
tion spectra in 2000. White barsdabsorption by CDOM; gray
barsdabsorption by phytoplankton; black barsdabsorption by non-
algal particulates. Small squares are absorption by CDOM, ag(440),
measured on grab samples. (b) Weekly averages of scattering
coe?cients at 555 nm (gray bars) in the Rhode River, and weekly
averages of the spectral exponent, h, characterizing the spectral
variability in scattering coe?cients (open circles) for 2000. Error bars
are of 95% con?dence limits.
C.L. Gallegos et al. / Estuarine, Coastin absorption in non-algal particulate matter at thetermination of the bloom that kept total absorption
coe?cient at bloom levels for longer than that due to the
phytoplankton bloomalone (Gallegos and Jordan, 2002).
The relative minimum in the sum of absorption
components in early summer, ca. day 160, following the
collapse of the bloom was much more noticeable in 2000
than in 2001, due to the much larger bloom that year.
Nevertheless, such a decline that also occurred in 2001
is borne out by measurements of scattering coe?cients
(see below).
4.3. Scattering coe?cients
The seasonal pattern of scattering coe?cient at
555 nm (Fig. 2b and Fig. 3b) was similar to that of
absorption by non-algal particulate matter. There was
a relative peak near the end of the spring phytoplankton
bloom (ca. day 135), a relative minimum near day 150
that was particularly evident in 2001, a gradual rise to
a seasonal maximum in late summer (ca. day 240), and
a rapid decline in fall (after day 270, only measured in
0
2
4
6
8
10
12
14
16
18
20
22
0.93
-0.71
Day of Year 2001
0.0
0.2
0.4
0.6
0.8
1.0
120 150 180 210 240 270 300 330 3600 30 60 90
120 150 180 210 240 270 300 330 3600 30 60 90
0
1
2
3
4
5
6
7
8
a
b
J MF M J S NA J A O D
J MF M J S NA J A O D
A
bs
or
pt
io
n 
C
oe
ffi
ci
en
t (
m
-1
)
b
p
(5
55
) 
(m
-1
)
-0.2
-0.4
-0.6
-0.8
-1.0
Fig. 3. As Fig. 2 for year 2001. Arrows and numbers designate points
chosen to demonstrate di?erences in scattering spectrum with di?erent
exponents (see Fig. 4).
161nd Shelf Science 64 (2005) 156e1702000). Scattering coe?cients were higher in 2000 than in
al a2001, until about mid-summer (i.e. day 200, Fig. 2b and
Fig. 3b).
The spectral exponent of scattering, h, varied widely
within and between years (Fig. 2b and Fig. 3b). In 2000
(Fig. 2b, open circles) there appeared to be a trend
characterized by relatively high values in early spring
soon after the breakup of ice (ca. days 60e90), declines
during the bloom of Prorocentrum minimum (days 110e
150), followed by a steady rise after day 150 to a late
summer peak. The most consistent feature between the 2
years was the relatively high values in late summer (days
200e270). The relatively low values during the smaller
spring bloom of P. minimum from ca. days 110e150 in
2001 were similar to those in 2000, but were not as
clearly associated with the bloom due to other low
values early in the year (Fig. 3b). The rise to the late
summer maximum in 2001 was interrupted by a sudden
decline to the lowest weekly averaged value observed in
the data set at about day 185. Examination of ancillary
data (Gallegos, unpublished) indicated that this time
period corresponded to a modest bloom of the small
dino?agellate Karlodinium micrum cf. of about 104
cells ml1, representing about a three-fold increase in
chlorophyll from about 15 to 45 mgm3. The largest
absorption due to phytoplankton observed during 2001
also occurred during this week (cf. Fig. 3a, ca. day 185).
Within 2 weeks the averaged h had returned to values
O0.6 (Fig. 3b), characteristic of late summer conditions
seen in 2000.
Examples of scattering spectra with low (0.71) and
high (0.93) values of h are given in Fig. 4. Depression of
scattering in the chlorophyll absorption bands in the blue
(412e488 nm) and the red (676 nm) was evident in the
spectrum with the low h (Fig. 4, ?lled squares), indicating
a noticeable e?ect of phytoplankton absorption (Stram-
ski and Mobley, 1997). The spectrum with the large h
(Fig. 4, open circles) decreased monotonically with
increasing wavelength, indicating dominance of the
scattering by smaller plankton or mineral particulates
(Babin et al., 2001).
4.4. Sources of materials and environmental covariates
Seasonal changes in the parameters that determine
optical properties are examined in relation to driving
processes for 2000, the year of more complete measure-
ments. Sources of CDOM include stream and ground-
water inputs from the local watershed and wetlands of
Muddy Creek, phytoplankton excretion and lysis, and
benthic decomposition of settled phytoplankton and
organic matter. Sinks for CDOM include bacterial
and photo-degradation. Exchange at the mouth with
Chesapeake Bay may be either a source or sink,
depending on the direction of the gradient.
The potential for freshwater inputs of CDOM peaked
162 C.L. Gallegos et al. / Estuarine, Coastduring the months of March through May whendischarge from both the local watershed of Muddy
Creek and from the Susquehanna River (the main
freshwater source to upper Chesapeake Bay) was
sustained at high levels (Fig. 5a). If freshwater inputs
were the main source of CDOM, then we would expect
CDOM absorption to vary inversely with salinity.
However, salinity was not a good correlate of CDOM
absorption, because salinity reached a minimum (ca. day
120) and increased steadily for most of the remainder of
the year (Fig. 5b, ?lled squares), while CDOM
absorption increased irregularly to ca. day 270, then
declined abruptly while salinity continued to increase
(cf. Fig. 5b and c).
The overall seasonal pattern of CDOM absorption
followed that of temperature (cf. Fig. 5b and c), es-
pecially in regard to the rapid decline in fall. Spatial
patterns varied seasonally, with minimal spatial gra-
dients in spring (with the exception of elevated concen-
trations at the mud?at station due to local ?ow ca. day
80) and late fall, and pronounced spatial gradients from
late spring through late summer (Fig. 5c). Only in late
winter (ca. day 60) during peak ?ow of the Susquehanna
River and in late spring (ca. day 150) following the
collapse of the widespread bloom of Prorocentrum
minimum (Gallegos and Jordan, 2002) did CDOM
absorption at the mouth of the Rhode River exceed
that at up-estuary stations (Fig. 5c). Otherwise CDOM
absorption increased in the up-estuary direction. It
would thus appear that in situ production of CDOM
dominated the seasonal signal, with production rates
increasing up-estuary parallel with the overall trophic
400 450 500 550 600 650 700 750
6
8
10
12
14
16
Wavelength (nm)
= 0.93
=-0.71
S
ca
tte
rin
g 
C
oe
ffi
ci
en
t (
m
-1
)
Fig. 4. Model ?ts to weekly averaged scattering spectra representing
the highest (open circles and dashed line, hZ 0.93) and lowest (?lled
squares and solid line, hZ 0.71) spectral exponents estimated in the
time series. In the case of low h, the model does not ?t the depression
of scattering observed in chlorophyll absorption bands, especially at
676 nm.
nd Shelf Science 64 (2005) 156e170gradient in the system (Jordan et al., 1991).
1000
na Fl
o1E-3
l aAbsorption by phytoplankton at 676 nm as estimated
from the monitoring data closely paralleled changes
in extracted chlorophyll (Fig. 6a). The seasonal vari-
ability in chlorophyll absorption was dominated by the
extraordinary spring bloom of Prorocentrum minimum,
described previously by Gallegos and Jordan (2002),
freshet of the Susquehanna River. A brief spike in
measured nitrate concentration to O12 mM can be seen
prior to the spring bloom (Fig. 6b), with concentrations
remaining mostly below 4 mM the remainder of the year.
Variations in chlorophyll absorption (or chlorophyll
concentration) bore no obvious relationship with any
0
2
4
6
8
10
12
14
16
18
S
al
in
ity
0
4
8
12
16
20
24
28
32
 T
em
pe
ra
tu
re
 (?
C
)
 Temp
120 150 180 210 240 270 300 330 3600 30 60 90
120 150 180 210 240 270 300 330 3600 30 60 90
120 150 180 210 240 270 300 330 3600 30 60 90
100
 Susquehanna
 Weir 102
0.0
0.5
1.0
1.5
2.0
 Monitoring site
 Mouth
 Mudflat
b
c
J M M A O DF A J J S N
J M M A O DF A J J S N
S
us
qu
eh
an
 S
tre
am
 
1E-4
1E-5
 Salinity
a
g
(4
40
) (
m
-1
)
Day of Year, 2000
Fig. 5. Time series of the measured concentration of CDOM and potential environmental correlates for year 2000: (a) freshwater ?ow by (solid line)
the North Branch of Muddy Creek measured at weir 102, and (dotted line) the Susquehanna River, main freshwater source for northern Chesapeake
Bay; (b) salinity (?lled squares) and temperature (open circles); (c) concentration of CDOM measured in grab samples from di?erent locations (see
Fig. 1) along the axis of the estuary, (?lled squares) the monitoring site, (open circles) mouth of the Rhode River, and (?lled triangle) the subtidal
mud?ats (up-estuary).10000
a
J M MF A J
 F
lo
w
 (m
3  
s-
1 )
C.L. Gallegos et al. / Estuarine, Coastawho observed the onset of the bloom in response to the0.01
0.1
1
A O DJ S N
w
 (m
3  
s-
1 )
163nd Shelf Science 64 (2005) 156e170other measured nutrient (cf. Fig. 6a and b).
80
 PO NO
al aAbsorption by non-algal particulate matter was
higher and more variable in late winter, i.e. days 1e75,
than in early winter (day 300e360) due to intermittent
disturbance of the shoreline by tidal movement of
ice (Fig. 7a). Samples collected at these times of ice
movement were grab samples taken o? the end of the
dock where the ac9 was housed, since the monitor could
not be run under those conditions. Non-algal particulate
absorption at 440 nm increased from about 1 m1 during
late spring to O3 m1 at the termination of the spring
bloom of Prorocentrum minimum, declined near day 150,
then rose steadily to a mid- to late summer maximum,
and declined rapidly in the fall back to ca. 1 m1
(Fig. 7a). The seasonal trend corresponded only weakly
to the variation in concentration of total suspended
solids (TSS), due to the extremely high concentration of
TSS measured during the spring bloom (Fig. 7b). The
partitioning inversion correctly attributed the absorption
wise, TSS concentration rose to a local peak in late
summer (ca. day 220) and declined steadily in the fall,
similar to absorption by non-algal particulate matter.
The overall seasonal trend in scattering coe?cient
(Fig. 7a, open circles) corresponded better with TSS
concentration than did ap  f(440), including the peak in
TSS concentration that occurred due to the spring
phytoplankton bloom. Nevertheless, the late summer
relative maxima and fall declines in both scattering
coe?cients and absorption by non-algal particulates
appeared to be governed by the variation of TSS
(Fig. 7).
4.5. Comparisons with long-term determinants of TSS
We examined average seasonal patterns in various
long-term data from the Rhode River system to try to
determine causal mechanisms regulating concentrations
0
2
4
6
8
10
12
14
16
18
20
0.0
0.2
0.4
0.6
0.8
120 150 180 210 240 270 300 330 36030 60 900
b
N
O
3 
an
d 
N
H
4 
(?
M
)
 Day of Year 2000
 P
O
4 
(?
M
)
43
 NH4
Fig. 6. Weekly averages of measured absorption by phytoplankton at 676 nm and potential environmental correlates for the year 2000: (a)
comparison of (?lled squares) measured af(676) with (open circles) measured chlorophyll concentration, and (b) measured concentrations of
macronutrients, (?lled squares) nitrate, (plus signs) ammonium, and (open circles) phosphate.22
120 150 130 60 900
0.0
0.5
1.0
1.5
2.0
2.5
3.0
a
a
 
(6
76
) (
m
-1
)
J M MF A J
J M MF A J
a (676)
164 C.L. Gallegos et al. / Estuarine, Coastby the spring bloom to phytoplankton (Fig. 6a). Other-1.0
210 240 270 300 330 360
0
50
100
150
200
250
S NJ A O D
S NJ A O D
 C
hl
or
op
hy
ll 
a
 (m
g 
m
-3
)
 Chl a
nd Shelf Science 64 (2005) 156e170of non-algal particulate matter. The average weekly
0J A JF M M
l aconcentration of TSS in long-term data from the region
around the monitoring station on the Rhode River has
peaks in spring corresponding with the average timing
of the spring phytoplankton bloom, and in late summer
(Fig. 8a, squares) coinciding with the timing of the late
summer maximum in ap  f(440) and scattering co-
e?cient (Fig. 7). The fraction of TSS that is not
combustible at 500 
C varied between 0.5 and 0.8, and
seemed to be only slightly lower during the phytoplank-
ton growing season from days 120 to 270 than during
other times (Fig. 8a, thin line). Consequently, estimates
of the organic and inorganic fractions of TSS exhibit
similar seasonal patterns (Fig. 8b and c). The seasonal
pattern of estimated dry weight of phytoplankton plus
ciliates follows the overall trend of the estimated organic
solids, and, excepting the overestimate of the phyto-
plankton contribution to dry weight during the spring
the organic solids concentration (Fig. 8b). Ciliate
biomass only adds noticeably to the dry weight of
phytoplankton during mid-summer. The primary di?er-
ence between the patterns of inorganic and organic
suspended solids is the presence of a broad peak in
inorganic suspended solids in early spring coinciding
with maximal watershed inputs and seasonal mean
windspeed (Fig. 8c).
The overall similarity of both the combustible and
the non-combustible fractions of TSS to the annual
temperature cycle (cf. Figs. 8 and 5b), resulting in a mid-
summer peak that is out of phase with watershed
delivery and windspeed, strongly suggests a dependence
of TSS concentration on biological activity. Not
surprisingly, therefore, the mid-summer peak of TSS
and its rapid decline in the fall parallel the long-term
average biomass of mobile benthic predators such
4
8
12
16
20
24
28
32
36
40
44
b
240 270 3000 30 60 90 120 150 180 210 360330
To
ta
l S
us
pe
nd
ed
 S
ol
id
s 
(g
 m
-3
)
Day of Year
Fig. 7. Weekly averages of (a) absorption by non-algal particulates, and (b) time series of suspended particulate matter concentration measured in
independent transect cruises for the year 2000. Large peak in TSS concentration in spring was due to the spring dino?agellate bloom, and precedes
a peak in non-algal particulate absorption.0 30 60 90 120 150 18
1
2
3
4
a
P
ar
tic
ul
at
e 
A
bs
or
pt
io
n 
(m
-1
)
J A JF M M
a
p-  
(440)
C.L. Gallegos et al. / Estuarine, Coastabloom, quantitatively accounts for a large proportion of240 270 300210 360
2
4
6
8
10
12
14
16
18
20
A O DJ S N
A O DJ S N
 P
ar
tic
ul
at
e 
S
ca
tte
rin
g 
(m
-1
)
330
b
p
(555)
165nd Shelf Science 64 (2005) 156e170as Calinectes sapidus (blue crabs) and a variety of ?sh
166 C.L. Gallegos et al. / Estuarine, Coastal and Shelf Science 64 (2005) 156e170180 210 240 270 300 330 36030 60 900 120 150
180 210 240 270 300 330 36030 60 900 120 150
180 210 240 270 300 330 36030 60 900 120 150
0
5
10
15
20
25
30
 TSS
0.0
0.2
0.4
0.6
0.8
1.0
 F
ra
ct
io
n 
In
or
ga
ni
c
 Inorganic 
fraction
0
2
4
6
8
10
12
0
2
4
6
8
10
12
14
16
18
20
22
 Inorganic soids
 Windspeed
 Epibenthic feeders
0
20
40
60
80
 S
ed
im
en
t F
lu
x 
(k
g 
ha
-1
 w
k-
1 )
a
b
c
To
ta
l S
us
pe
nd
ed
 S
ol
id
s 
(g
 m
-3
)
J M J J S NF A M A O D
J M J J S NF A M A O D
J M J J S NF A M A O D
D
ry
 W
ei
gh
t (
g 
m
-3
)
100
IS
S
 (g
 m
-3
); 
 V
 (k
m
 h
-1
)
B
io
m
as
s 
(k
g 
tra
w
l-1
)
Day of Year
Organic solids
Phytoplankton
Phyto+Ciliates
 Sediment Flux
Fig. 8. Time series of weekly (unless otherwise speci?ed) medians of long-term measurements of suspended particulate matter and related parameters
considered potential drivers of the average seasonal pattern. (a) Concentration of (squares) total suspended particulate matter, and (thin line) fraction
of TSS that is inorganic (monthly data). Error bars are G1 standard error; (b) dry weight of (?lled squares) phytoplankton estimated from
chlorophyll concentration and (open circles) phytoplankton plus ciliates, compared with (gray bars) estimated organic fraction of TSS. Ciliates add
measurably to total dry weight in mid-summer. Sum of phytoplankton and ciliates accounts for most of the variation in organic suspended solids; (c)
seasonal pattern of (gray bars) estimated inorganic component of TSS compared with factors that input or resuspend inorganic sedimentsd(bold
line) windspeed, (bold dashed line) watershed in?ux from stream?ow, and (thin line and squares) biomass of epibenthic predators (monthly data).
Epibenthic predators consist mainly of croaker (Micropogonias undulates, ca. 2e3% of summer peak), spot (Leiostomus xanthurus, ca. 20% of
summer peak), and blue crabs (Callinectes sapidus, ca. 76e78% of summer peak).
al a(Fig. 8c, squares) which could contribute to sediment
resuspension by their feeding activity.
5. Discussion
Together the phytoplankton and non-algal particu-
late matter dominated the variability in optically active
constituents in the Rhode River (Figs. 2 and 3).
Interannual variability in the magnitude of the spring
phytoplankton bloom was pronounceddthe 2 years
observed here included blooms of extraordinary (2000)
and average (2001) magnitudes for this system (Gallegos
et al., 1997). Modeling studies (Gallegos et al., 1997)
and detailed automated monitoring data (Gallegos and
Jordan, 2002) have shown that interannual variability in
the magnitude of the spring bloom in the Rhode River is
governed by the timing and magnitude of nitrate inputs
by the spring freshet of the Susquehanna River. The
freshet must be large enough to carry high concen-
trations of nitrate, must not arrive prior to adequate
daylength and seasonal phosphate availability (i.e. after
mid-March), or so late that the nitrate is consumed
during its transit along the axis of northern Chesapeake
Bay (i.e. before mid-May) (Gallegos et al., 1997;
Gallegos and Jordan, 2002).
Absorption by CDOM in summer 2000 also exceeded
that in 2001 (Fig. 2a and Fig. 3a). If the interannual
di?erence in CDOM absorption was related to the
di?erences in relative magnitudes of the spring phyto-
plankton blooms, then the increase took longer to be
expressed than the absorption by non-algal particulates.
The observations are consistent with death of autotrophs
and growth of heterotrophic plankton producing the
peak in non-algal absorption following bloom collapse,
and longer-term decomposition over the summer pro-
ducing the higher levels of absorption by CDOM. The
lack of correlation between CDOM absorption and
salinity di?ers from some studies that have examined
spatial gradients in coastal systems in individual cruises
(Twardowski and Donaghay, 2001; Hu et al., 2003).
However, because the Rhode River is a subestuary,
salinity at the mouth varied from ca. 5 to 16 over the
observation period (similar to that at the monitoring site,
Fig. 5b). Within this restricted salinity range, Harding
et al. (2005) observed CDOM absorption in the main
stem of Chesapeake Bay varying from 0.3 to 0.8 m1,
and only weakly correlated with salinity. Similarly,
CDOM absorption of local stream ?ow varied from
!0.5 to O3 m1 (Gallegos, unpublished), so that weak
correlation between CDOM absorption and salinity on
seasonal time scales is not unexpected. Gallegos (2005)
observed similar lack of correlation between salinity and
CDOM absorption in the humic-stained St. Johns River,
C.L. Gallegos et al. / Estuarine, Coastdue primarily to large spatial and temporal variations inCDOM absorption at zero-salinity within the tidal
freshwater portion of the estuary.
The interannual variability in non-algal particulate
matter appeared less pronounced than that of phyto-
plankton, except for the peak in non-algal particulate
absorption associated with the collapse of the extraordi-
nary spring phytoplankton bloom in 2000 which was not
observed in 2001. The mid- to late summer peaks in
scattering coe?cient and non-algal particulate absorp-
tion, as well as the fall declines were similar in the 2 years
(Fig. 2b andFig. 3b). The late summer increase in spectral
exponent of scattering (Fig. 2b and Fig. 3b) indicates
a decrease in the particle-size spectrum consistent with
known seasonal shifts in species composition from
diatoms to picoplankton and ?agellates in Chesapeake
Bay (Malone et al., 1996), and/or increasing mineral
contribution to scattering (Babin et al., 2003).
As was observed on the continental shelf of the
Middle Atlantic Bight (Chang and Dickey, 2001), much
of the variability in optical properties in the Rhode River
subestuary was related to season. By examining weekly
averages of optical properties, we removed much of the
variability in our data associated with short-term
episodic events that characterized the observations of
Chang and Dickey (2001) on the continental shelf. In this
study, however, we were also able to observe interannual
variability, which was pronounced with respect to the
magnitude of the spring phytoplankton bloom and its
collapse. The contrasting years help us resolve variability
that was directly related to phytoplankton blooms from
that which was not. Aside from the di?erences in the
magnitude of phytoplankton absorption, the largest
di?erence between the 2 years was the pronounced peak
in absorption by non-algal particulates as the extraordi-
nary bloom collapsed in 2000, which was missing in 2001.
Gallegos and Jordan (2002) showed that the increase in
non-algal absorption in 2000 initially coincided with
a transient increase in empty thecae of Prorocentrum
minimum, indicating that cell mortality was a likely
factor at ?rst. Growth of bacteria and other heterotro-
phic plankton may have sustained the peak in non-algal
absorption. The observation that the late summer
maximum in non-algal particulate absorption (Fig. 2a
and Fig. 3a) and scattering coe?cient (Fig. 2a and
Fig. 3a) were similar between the 2 years indicates that
the formation of the late summer maximum in particu-
late matter does not depend upon having an extraordi-
nary phytoplankton bloom in the spring.
The late summer peak in absorption by non-algal
particulate matter appeared to have a biological expla-
nation (Fig. 8c). As is common in many low energy
coastal systems (Pempkowiak et al., 2002), a benthic
?u? layer forms on the bottom of the Rhode River in
the region of the optical monitoring station, and is
especially pronounced during the summer. Where they
167nd Shelf Science 64 (2005) 156e170have been studied, such layers have been shown to be
l ahighly ?uidized (Amos et al., 1997), composed of
a mixture of sedimented phytoplankton (Beaulieu,
2003), protists (Shimeta et al., 2002), and ?ne sediment
(Pempkowiak et al., 2002), and resuspended much more
easily than consolidated sediments of similar composi-
tion (Maa et al., 1998; Leipe et al., 2000). Increased
erodability of this material even under the reduced
winds of summer might be su?cient to produce the mid-
summer peak in TSS concentration. Hines et al. (1990)
demonstrated that bioturbation by the epibenthic
predatory guild of blue crabs and ?sh disturbed the
top layer of benthic sediments to a depth of 10 cm at
a station in the Rhode River during summer. The
activity of these organisms may, therefore, interact with
wind and tidal resuspension to produce the mid-summer
peak in both organic and inorganic suspended particu-
late matter. Biological activity was also shown to be
a factor in resuspension of a benthic ?u? layer that
varied in amount with season and location in lower
Chesapeake Bay (Maa and Lee, 1997). Similarly,
epibenthic ?sh were shown to facilitate sediment
resuspension, while experimental removal of ?sh pro-
moted bottom sediment consolidation in a shallow lake
in the Netherlands (Sche?er et al., 2003).
Additionally, seasonal formation of the benthic ?u?
layer may, in part, rely on particle ?occulation and
sedimentation of ?ocs from the water column. Though
salt ?occulation of charged primary clay particles is
rapid, formation of macro?ocs consisting of coated
minerals and detritus is enhanced by polysaccharides
and polymers produced by bacteria (van Leussen, 1988).
Such a process might scavenge the water column of
particles as temperature rises and the spring bloom
collapses in late spring (Jones et al., 1998), possibly
accounting for the late spring minimum in inorganic
suspended solids (Fig. 8c) and scattering coe?cients
(Fig. 2b, Fig. 3b and Fig. 7) around day 150.
These results have implications for the development
of indicators of water quality and habitat suitability for
living resources in the coastal zone. Gallegos and Jordan
(2002) demonstrated that absorption and scattering by
non-algal particulate matter extended the impact of an
extraordinary dino?agellate bloom on light attenuation
an additional 2 weeks longer than that due to phyto-
plankton alone. Furthermore, if we are correct that
biotic resuspension of a phytodetritus and sediment ?u?
layer is responsible for the late summer maximum in
attenuation, then it is clear that eutrophication has long-
term e?ects on optical properties of a water body beyond
the immediate absorption and scattering of light directly
attributable to phytoplankton. Neither chlorophyll
concentration nor phytoplankton species composition
provides a measure of this aspect of eutrophication.
Finally, the independence of this process from the
magnitude of the preceding spring bloom indicates that
168 C.L. Gallegos et al. / Estuarine, Coastathe e?ects of eutrophication on optical properties ofthe non-algal particulate matter may take multiple years
to reverse.
Acknowledgments
Support for this work was provided by the U.S.
Environmental Protection Agency through the Coastal
Intensive Site Network (CISNet) grant R826943 and the
Estuary and Great Lakes Coastal Initiative (EaGLes)
grant R82868401, and by the Smithsonian Environmen-
tal Sciences Program. The research described in this
paper has not been subjected to the U.S. Environmental
Agency?s required peer and policy review and therefore
does not necessarily re?ect the views of the Agency and
no o?cial endorsement should be inferred. We thank
D.W. Coats for data on ciliate biovolume, and K. Yee,
D. Sparks, J. Miklas, and R. Aguilar for help with data
collection and analyses.
Appendix A
The ac9 measures absorption coe?cients referenced
to the absorption by pure water, which is subtracted by
the manufacturer?s software. In addition, the measured
coe?cients overestimate true absorption due to loss of
photons by scattering within the re?ective tube, which
the instrument incorrectly attributes to absorption (Kirk,
1992). Kirk (1992) demonstrated that, for a given
scattering phase function, the error is proportional to
the scattering coe?cient, which can be used to derive
a correction factor (Zaneveld et al., 1994; Gallegos and
Neale, 2002),
atw?l?Zam?l?  3?c?l?  am?l?Zam?l?  3bm?l? ?A1?
where l is the wavelength of light, at  w(l) is the true
total absorption less that due to water, am(l) is the
absorption coe?cient measured by the ac9, c(l) is the
attenuation coe?cient, bm(l)hc(l)  am(l) is the mea-
sured scattering coe?cient, and 3 is a correction factor.
We estimated 3 by the procedure of Gallegos and Neale
(2002) using their unconstrained matrix inversion. Note
that bm(l) underestimates the true particulate scattering,
bp(l), by the factor (1C 3).
We estimated the components of absorption by
a modi?cation of the procedure of Gallegos and Neale
(2002). We represented the absorption coe?cient
referenced to water as the sum of absorption due to
colored dissolved organic matter (CDOM), phytoplank-
ton pigments, and non-algal particulate matter,
atw?l?Zag?l?Caf?l?Capf?l? ?A2?
where the subscripts refer to absorption by particular
nd Shelf Science 64 (2005) 156e170components: g for CDOM (i.e. gelbsto?, Kirk, 1994),
al af for phytoplankton, and p  f for non-algal particulate
matter. The component absorption spectra were repre-
sented by normalized absorption functions, scaled by
the absorption at a characteristic wavelength,
ag?l?Zag?440?g?l? ?A3a?
af?l?Zaf?676?f?l? ?A3b?
and
apf?l?Zapf?440?p?l? ?A3c?
where g(l), f(l), and p(l) are the normalized absorp-
tion functions for, respectively, CDOM, phytoplankton,
and non-algal particulate matter. Substituting Eqs.
(A3a)e(A3c) into Eq. (A2) and equating to the right
hand side of Eq. (A1) produces a linear system of
equations with 4 unknowns consisting of the scale
factors in Eqs. (A3a)e(A3c) and 3, measurements am(l)
and bm(l), and coe?cients given by the normalized
absorption spectra. The system is fully determined using
measurements at 4 wavelengths, so that the scale factors
and 3 may be determined by matrix inversion. Gallegos
and Neale (2002) called this procedure the uncon-
strained solution.
The unconstrained solution performed well for
estimating af(676) and 3, but sometimes produced
unstable estimates of ag(440) and ap  f(440), due to
the similar shapes for g(l) and p(l). Gallegos and Neale
(2002) used a statistically augmented procedure to
improve the discrimination between absorption by
CDOM from that by non-algal particulates, by noting
that non-algal particulate matter contributes to scatter-
ing, whereas CDOM does not. They parameterized the
scale factor for absorption by non-algal particulates in
terms of the absorption-to-scattering ratio of non-algal
particulates at 440 nm, r, where
rhapf?440?
bp?440?
?A4?
and estimated r by a site-speci?c statistical model.
However, over or underestimates of r could still produce
negative estimates of ag(440) or ap  f(440). Here we
further modi?ed the procedure by noting that r must lie
along the line,
rZrcm?1 fc? ?A5?
where rcmZ

apf?440?Cag?440?

=bp?440? is the ab-
sorption-to-scattering ratio of particulate plus dissolved
matter and is ?xed by the measurements once 3 and
af(676) have been estimated, and fcZag?440?=
apf?440?Cag?440?

is the fraction of total non-algal
absorption at 440 nm attributable to CDOM. We used
the statistically augmented procedure of Gallegos and
C.L. Gallegos et al. / Estuarine, CoastNeale (2002) to provide a ?rst estimate of r^, a^g?440?,and f^c, then revised the estimate by ?nding the nearest
point on Eq. (A5) that was also on the normal to Eq.
(A5), passing through (f^c, r^). The revised procedure
guarantees positive estimates for ag(440) and r whenever
measured spectra conform to the shape predicted by
Eqs. (A2), (A3a)e(A3c). Note that, though the pro-
cedure only estimates 3 scale coe?cients at a character-
istic wavelength, the procedure exploits the spectral
variability in absorption at 4 widely spaced wavelengths.
The complete spectral shape of absorption is implied by
the scale factors applied to the normalized absorption
functions. It is, of course, not possible to simultaneously
estimate the scale factors and the normalized absorption
functions.
References
Aksnes, D.L., Nejstgaard, J., Soedberg, E., S?rnes, T., 2004. Optical
control of ?sh and zooplankton populations. Limnology and
Oceanography 49, 233e238.
Amos, C.L., Feeney, T., Sutherland, T.F., Luternauer, J.L., 1997. The
stability of ?ne-grained sediments from the Fraser River Delta.
Estuarine, Coastal and Shelf Science 45, 507e524.
Babin, M., Morel, A., Fournier-Sicre, V., Fell, F., Stramski, D., 2003.
Light scattering properties of marine particles in coastal and open
ocean waters as related to the particle mass concentration.
Limnology and Oceanography 48, 843e859.
Beaulieu, S.E., 2003. Resuspension of phytodetritus from the sea ?oor:
a laboratory ?ume study. Limnology and Oceanography 48,
1235e1244.
Bockstahler, K.R., Coats, D.W., 1993. Grazing of the mixotrophic
dino?agellate Gymnodinium sanguineum on ciliate populations of
Chesapeake Bay. Marine Biology 116, 477e487.
Chang, G.C., Dickey, T.D., 1999. Partitioning in situ total spectral
absorption by use of moored spectral absorptioneattenuation
meters. Applied Optics 38, 3876e3887.
Chang, G.C., Dickey, T.D., 2001. Optical and physical variability on
timescales from minutes to the seasonal cycle on the New England
shelf: July 1996 to June 1997. Journal of Geophysical Research 106,
9434e9453.
Coats, D.W., Heinbokel, F.J., 1982. A study of reproduction and other
life cycle phenomena in planktonic protists using an acridine
orange ?uorescence technique. Marine Biology 67, 71e79.
Correll, D.L., 1981. Nutrient mass balances for the watershed,
headwaters, intertidal zone, and basin of the Rhode River estuary.
Limnology and Oceanography 26, 1142e1149.
Correll, D.L., Jordan, T.E., Weller, D.E., 1999. Precipitation e?ects on
sediment and associated nutrient discharges from Rhode River
watersheds. Journal of Environmental Quality 28, 1897e1907.
Davies-Colley, R.J., Vant, W.N., Smith, D.G., 1993. Colour and
clarity of natural waters. Ellis Horwood.
Dawson, C.E., 1965. Lengtheweight relationships of some Gulf of
Mexico ?shes. Transactions of the American Fisheries Society 94,
279e280.
Dickey, T., Frye, D., Jannasch, H., Boyle, E., Manov, D.,
Sigurdson, D., McNeil, J., Stramska, M., Michaels, A.,
Nelson, N., Siegel, D., Chang, G., Wu, J., Knap, A., 1998. Initial
results from the Bermuda testbed mooring program. Deep-Sea
Research I 45, 771e794.
Fenchel, T., Finlay, B.J., 1983. Respiration rates in heterotrophic, free-
living protozoa. Microbial Ecology 9, 99e122.
Gallegos, C.L., 2001. Calculating optical water quality targets to
169nd Shelf Science 64 (2005) 156e170restore and protect submersed aquatic vegetation: overcoming
problems in partitioning the di?use attenuation coe?cient for Maa, J.P.-Y., Sanford, L.P., Halka, J.P., 1998. Sediment resuspension
170 C.L. Gallegos et al. / Estuarine, Coastal and Shelf Science 64 (2005) 156e170photosynthetically active radiation. Estuaries 24, 381e397.
Gallegos, C.L., 2005. Optical water quality of a blackwater river
estuary: the Lower St. Johns River, Florida, USA. Estuarine,
Coastal and Shelf Science 63, 57e72.
Gallegos, C.L., Jordan, T.E., 2002. Impact of the Spring 2000
phytoplankton bloom in Chesapeake Bay on optical properties
and light penetration in the Rhode River, Maryland. Estuaries 25,
508e518.
Gallegos, C.L., Jordan, T.E., Correll, D.L., 1992. Event-scale response
of phytoplankton to watershed inputs in a subestuary: timing,
magnitude, and location of phytoplankton blooms. Limnology and
Oceanography 37, 813e828.
Gallegos, C.L., Kenworthy, W.J., 1996. Seagrass depth limits in the
Indian River Lagoon (Florida, USA): application of an optical
water quality model. Estuarine, Coastal and Shelf Science 42,
267e288.
Gallegos, C.L., Neale, P.J., 2002. Partitioning spectral absorption
in case 2 waters: discrimination of dissolved and particulate
components. Applied Optics 41, 4220e4233.
Gallegos, C.L., Correll, D.L., Pierce, J.W., 1990. Modeling spectral
di?use attenuation, absorption, and scattering coe?cients in a
turbid estuary. Limnology and Oceanography 35, 1486e1502.
Gallegos, C.L., Jordan, T.E., Correll, D.E., 1997. Interannual
variability in spring bloom timing and magnitude in the Rhode
River, Maryland (USA): observations and modeling. Marine
Ecology Progress Series 154, 27e40.
Harding, L.W.J., Magnuson, A., Mallonee, M.E., 2005. SeaWiFS
retrievals of chlorophyll in Chesapeake Bay and the mid-Atlantic
bight. Estuarine, Coastal and Shelf Science 62, 75e94.
Hines, A.H., Haddon, A.M., Wiechert, L.A., 1990. Guild structure
and foraging impact of blue crabs and epibenthic ?sh in
a subestuary of Chesapeake Bay. Marine Ecology Progress Series
67, 105e126.
Hu, C., Muller-Karger, F.E., Biggs, D.C., Carder, K.L., Nababan, B.,
Nadeau, D., Vanderbloemen, J., 2003. Comparison of ship and
satellite bio-optical measurements on the continental margin of the
NE Gulf of Mexico. International Journal of Remote Sensing 24,
2597e2612.
Jones, S.E., Jago, C.F., Bale, A.J., Chapman, D., Howland, R.J.M.,
Jackson, J., 1998. Aggregation and resuspension of suspended
particulate matter at a seasonally strati?ed site in the southern
North Sea: physical and biological controls. Continental Shelf
Research 18, 1283e1309.
Jordan, T.E., Pierce, J.W., Correll, D.L., 1986. Flux of particulate
matter in the tidal marshes and subtidal shallows of the Rhode
River estuary. Estuaries 9, 310e319.
Jordan, T.E., Correll, D.L., Miklas, J., Weller, D.E., 1991. Nutrients
and chlorophyll at the interface of a watershed and an estuary.
Limnology and Oceanography 36, 251e267.
Kirk, J.T.O., 1992. Monte Carlo modeling of the performance of
a re?ective tube absorption meter. Applied Optics 31, 6463e6468.
Kirk, J.T.O., 1994. Light and Photosynthesis in Aquatic Ecosystems,
second ed. Cambridge University Press.
Leipe, T., Loe?er, A., Emeis, K.-C., Jaehmlich, S., Bahlo, R.,
Ziervogel, K., 2000. Vertical patterns of suspended matter
characteristics along a coastal-basin transect in the western Baltic
Sea. Estuarine, Coastal and Shelf Science 51, 789e804.
van Leussen, W., 1988. Aggregation of particles, settling velocity of
mud ?ocs, a review. In: Dronkers, J., van Leussen, W. (Eds.),
Physical Processes in Estuaries. Springer-Verlag, pp. 347e403.
Maa, J.P.-Y., Lee, C.-H., 1997. Variation of the resuspension
coe?cients in the lower Chesapeake Bay. Journal of Coastal
Research 25, 63e74.characteristics in Baltimore Harbor, Maryland. Marine Geology
146, 137e145.
Malone, T.C., Conley,D.J., Fisher, T.R., Glibert, P.M.,Harding, L.W.,
Sellner, K.G., 1996. Scales of nutrient-limited phytoplankton
productivity in Chesapeake Bay. Estuaries 19, 371e385.
McPherson, B.F., Miller, R.L., 1987. The vertical attenuation of light
in Charlotte Harbor, a shallow, subtropical estuary, south-western
Florida. Estuarine, Coastal and Shelf Science 25, 721e737.
Mobley, C.D., 1994. Light and water. Radiative transfer in natural
waters. Academic Press, New York.
Montagnes, D.J.S., Lynn, D.H., 1993. A quantitative protargol stain
(QPS) for ciliates and other protists. In: Kemp, P.F., Sherr, B.F.,
Sherr, E.B., Cole, J.J. (Eds.), Handbook of Methods in Aquatic
Microbial Ecology. Lewis Publishers, pp. 229e240.
Olmi III, E.J., Bishop, J.M., 1983. Variations in total widtheweight
relationships of blue crabs, Callinectes sapidus, in relation to sex,
maturity, molt stage and carapace form. Journal of Crustacean
Biology 3, 575e581.
Pempkowiak, J., Beldowski, J., Pazdro, K., Staniszewski, A., Leipe, T.,
Emeis, K.-C., 2002. The contribution of the ?ne sediment fraction
to the ?u?y layer suspended matter (FLSM). Oceanologia 44,
513e527.
Postma, H., 1967. Sediment transport and sedimentation in the
estuarine environment. In: Lau?, G.H. (Ed.), Estuaries. American
Association for the Advancement of Science, Washington, DC, pp.
158e179.
Rabalais, N.N., Turner, R.E., 2001. Coastal hypoxia: consequences for
living resources and ecosystems. American Geophysical Union.
Richardson, K., J?rgensen, B.B., 1996. Eutrophication: de?nition,
history and e?ects. In: J?rgensen, B.B., Richardson, K.R. (Eds.),
Eutrophication in Coastal Marine Ecosystems. American
Geophysical Union, pp. 1e19.
Sanford, L.P., 1994. Wave-forced resuspension of upper Chesapeake
Bay muds. Estuaries 17, 148e165.
Sche?er, M., Portielje, R., Zambrano, L., 2003. Fish facilitate wave
resuspension of sediment. Limnology and Oceanography 48,
1920e1926.
Shimeta, J., Amos, C.L., Beaulieu, S.E., Ashiru, O.M., 2002.
Sequential resuspension of protists by accelerating tidal ?ow:
implications for community structure in the benthic boundary
layer. Limnology and Oceanography 47, 1152e1164.
Southwick, C.H., Pine, F.W., 1975. Abundance of submerged vascular
vegetation in the Rhode River from 1966 to 1973. Chesapeake
Science 16, 147e151.
Stramski, D., Mobley, C.D., 1997. E?ects of microbial particles on
oceanic optics: a database of single-particle optical properties.
Limnology and Oceanography 42, 538e549.
Stross, R.G., Sokol, R.C., 1989. Runo? and ?occulation modify
underwater light environment of the Hudson River estuary.
Estuarine, Coastal and Shelf Science 29, 305e316.
Stumpf, R.P., Pennock, J.R., 1989. Calibration of a general opti-
cal equation for remote sensing of suspended sediments in
a moderately turbid estuary. Journal of Geophysical Research 94,
14,363e14,371.
Stumpf, R.P., Pennock, J.R., 1991. Remote estimation of the di?use
attenuation coe?cient in a moderately turbid estuary. Remote
Sensing of Environment 38, 183e191.
Twardowski, M.S., Donaghay, P., 2001. Separating in situ terrigenous
sources of absorption by dissolved materials in coastal waters.
Journal of Geophysical Research 106, 2545e2560.
Zaneveld, J.R.V., Kitchen, J.C., Moore, C., 1994. The scattering error
correction of re?ecting-tube absorption meters. In: Ackleson, S.
(Ed.), Ocean Optics XII. SPIE, pp. 44e55.