592
Limnol. Oceanogr., 46(3), 2001, 592?603
q 2001, by the American Society of Limnology and Oceanography, Inc.
Ultraviolet radiation sensitivity of photosynthesis in phytoplankton from an estuarine
environment
Anastazia T. Banaszak1 and Patrick J. Neale2
Smithsonian Environmental Research Center, Edgewater, Maryland 21037
Abstract
We have studied temporal variation in the sensitivity of phytoplankton photosynthesis to inhibition by ultraviolet
radiation (UV; 280?400 nm) using biological weighting functions (BWFs) that quantify the biological effect of
different wavelengths of UV. Variations in irradiance-dependent BWFs were evaluated for natural phytoplankton
assemblages from the Rhode River, a shallow subestuary of the Chesapeake Bay, Maryland, from October 1994 to
July 1996. Phytoplankton assemblages were sensitive to UV throughout the year. Rhode River assemblages are
inhibited more strongly in the UV-B (280?320 nm), particularly below 300 nm, but there is a significant influence
well into the UV-A (320?400 nm). There was no inhibition of phytoplankton photosynthesis by photosynthetically
available radiation (400?700 nm), but there was significant seasonal variation in the saturated rate of photosynthesis
(P ) and in the light saturation parameter (Es). There was little variation in seasonal average BWFs through theBs
year, but there was considerable variation in BWFs during each season. Individual BWFs varied both in absolute
values of the weightings (reciprocal [mW m22]) and in the spectral shape or relative effect of UV-B versus UV-A,
which may be due to changes in species composition, light, temperature, and nutrient availability. Comparison of
the most sensitive assemblage (spring) with the least sensitive assemblage (winter) indicates that these BWFs are
close to the upper and lower bounds in sensitivity for irradiance-dependent BWFs from all natural and cultured
phytoplankton populations. The average, absolute spectral weightings for inhibition of photosynthesis in assemblages
from the Rhode River are similar to an average BWF for Antarctic assemblages.
Ultraviolet radiation (UV; 280?400 nm) has deleterious
effects on aquatic organisms that are manifested at genetic,
physiological, and biochemical levels. These effects occur in
habitats ranging from tropical environments, where UV ra-
diation is naturally high, to polar environments, where the
levels of UV radiation can become elevated during spring-
time ozone-depletion events. Stratospheric ozone loss results
in a wavelength-dependent shift of incident UV-B (280?320
nm; however, there is no significant surface irradiance for
wavelengths below 290 nm) because of the penetration of
shorter wavelengths, which are more energetic and more bi-
ologically harmful (Lubin et al. 1992). In comparison, UV-
A (320?400 nm) and photosynthetically available radiation
(PAR; 400?700 nm) remain essentially unaffected by ozone
depletion.
Assessments of the effects of variable UV-B have to be
made in the context of the full UV spectrum, because both
UV-B and UV-A can affect physiological processes and con-
1 Present address: Unidad Acade?mica?Puerto Morelos, Instituto
de Ciencias del Mar y Limnolog??a, Universidad Nacional Auto?noma
de Me?xico, Apartado Postal 1152, Cancu?n, 77500, Me?xico
2 Corresponding author.
Acknowledgments
We thank Sharyn Hedrick for help with fieldwork and Catherine
Jarriel for help with experiments. A.T.B. was funded by a Smith-
sonian Institution Postdoctoral Fellowship. Charles Gallegos and
Karen Yee are acknowledged for making available optical data from
the Rhode River, which were acquired with support from Environ-
mental Protection Agency, CISNET grant R826943. Additional sup-
port was provided by the National Science Foundation Office of
Polar Programs grant OPP-9615342. Charles Gallegos, Elena Litch-
man, and John Cullen provided comments that greatly improved the
manuscript. This paper is dedicated to the memory of Douglass
Hayes, Jr.
tribute to changes in community structure (Bothwell et al.
1994). Phytoplankton are constrained to photic environments
where they are exposed to PAR and UV, possibly requiring
a tradeoff between absorption of adequate PAR for photo-
synthesis and increased exposure to damaging UV. The in-
hibition of phytoplankton photosynthesis by UV is one effect
of major ecological significance, although it is not the only
type of damage that can occur (Vincent and Neale 2000 and
references therein). Numerous studies have shown that UV-
B and UV-A inhibit photosynthesis in natural populations
(Helbling et al. 1992; Smith et al. 1992; Behrenfeld et al.
1993; Neale et al. 1994) and cultures (Cullen and Lesser
1991; Cullen et al. 1992; Neale et al. 1998a).
To identify the physiological and ecological processes that
dominate phytoplankton responses to UV and to assess what
portion of responses is associated with ozone-dependent
changes in UV-B requires the development of wavelength-
dependent weighting functions. Biological weighting func-
tions (BWFs) describe the effectiveness of radiation of dif-
ferent wavelengths to produce a biological response such as
inhibition of photosynthesis. A wavelength-dependent de-
scription of UV effects on photosynthesis (the BWF/P-I
model) has been developed that describes photosynthesis as
a function of PAR and photoinhibition as a function of both
PAR and UV (Cullen et al. 1992; Neale et al. 1994). Two
versions of the BWF/P-I model (the E and H models) are in
use contingent on whether inhibition of photosynthesis is
dependent on the rate of exposure (E, irradiance) or cumu-
lative exposure (H). The irradiance-dependent model (the E
model) applies when active repair counterbalances the effect
of UV damage and photosynthesis attains a steady state dur-
ing exposure, as has been observed for cultures of temperate
phytoplankton (Lesser et al. 1994). Experimental exposures
used for estimating a BWF are polychromatic?i.e., shorter
593UV sensitivity of phytoplankton
Table 1. Dates and sites of collections of water samples for determination of biological weighting functions (BWFs) and the dominant
phytoplankton species present. Sample sites are described in Gallegos et al. (1992). Sta. 4B is located in the Rhode River, ;4 km upriver
from the mouth on the Chesapeake Bay. ??Dock?? refers to a sample taken at the same channel position as Sta. 4B but ;20 m from shore
at the end of the SERC pier; Sta. 1A is located 1.4 km east of the mouth of the Rhode River to the Chesapeake Bay; the Bay Bridge is
located on the Chesapeake Bay ;20 km north of the Rhode River mouth. ??Int.?? refers to integrated (whole water-column) samples, and
??Surf.?? refers to samples taken of surface water only.
Date Sampling site Dominant diatom Dominant dinoflagellate Other dominant flagellates
13 Oct 94
03 Nov 94
21 Nov 94
11 Jan 95
07 Mar 95
06 Apr 95
4B Int.
4B Int.
4B Int.
Dock Int.
4B Int.
4B Int.
Cyclotella glomerata
C. glomerata
Skeletonema costatum
Chaetoceros sp.
Chaetoceros sp.
Katodinium rotundatum
K. rotundatum
K. rotundatum
Cryptomonas sp.
Cryptomonas sp.
Cryptomonas sp.
18 Apr 95
03 May 95
16 May 95
31 May 95
01 Jun 95
12 Jul 95
4B Int.
4B Int.
4B Int.
Dock Int.
Dock Int.
Bay Bridge Surf.
Cerataulina pelagica
Thalassiosira pseudonana
T. pseudonana
Chaetoceros debilis
Prorocentrum minimum
P. minimum
Gymnodinium sp.
Gyrodinium uncatenum
G. uncatenum
K. rotundatum
Cryptomonas sp.
Chrysochromulina sp.
Chrysochromulina sp.
18 Jul 95
27 Jul 95
02 Aug 95
08 Aug 95
22 Aug 95
13 Sep 95
4B Int.
4B Int.
1A Surf.
4B Surf.
4B Surf.
4B Surf.
Chaetoceros sp.
Chaetoceros sp.
P. minimum
P. minimum
Scrippsiella trochoidea
S. trochoidea
Gymnodinium sp.
Polykrikos hartmanni
Chroomonas sp.
25 Oct 95
20 Nov 95
13 Dec 95
27 Mar 96
03 Jul 96
4B Int.
Dock Int.
4B Int.
Dock Surf.
Dock Int.
S. costatum
S. costatum
K. rotundatum
K. rotundatum
P. minimum
Cryptomonas sp.
Apedinella radians
wavelengths of UV-A then UV-B are added to a constant
background of PAR, as occurs during environmental expo-
sure (Rundel 1986). It is critical for BWFs to have both high
resolution and spectral accuracy in the UV-B region. Solar
spectral irradiance drops by three orders of magnitude or
more in the 290?320 nm region because of absorption by
stratospheric ozone. Small changes in absolute energy levels
may represent large effects when weighted by a BWF and
may substantially change predictions of UV-B?induced pho-
toinhibition.
The level of photoinhibition will be modified by many
factors that influence the extent of exposure and the sensi-
tivity (BWF) of the phytoplankton. High irradiance will en-
hance photoinhibition when the damage incurred exceeds the
capacity of photoprotective and repair processes (Neale
1987). Phytoplankton encountering shallow mixing condi-
tions (Neale et al. 1998c) or forming surface blooms will be
exposed to high irradiance. Photoinhibition will extend deep-
er in water bodies with low attenuation coefficients and
greater penetration of UV radiation.
Photoprotective and repair processes are particularly im-
portant in preventing and reversing the damage to photosyn-
thesis by UV (Lesser et al. 1994; Neale et al. 1994). The
induction of photoprotective mechanisms could change the
absolute weights and shapes of BWFs by the presence of
pigments or UV-screening compounds that absorb in the
wavelength range causing the damage (Neale et al. 1994).
A specific decrease in the weighting function in the region
of absorbance of these compounds would indicate a photo-
protective function, as was found for mycosporine-like ami-
no acids (MAAs) in Gymnodinium sanguineum (Neale et al.
1998a). Carotenoids also act in a photoprotective function
against high irradiance (Demers et al. 1991). Organisms have
varying abilities to reverse the effects of UV radiation (Kar-
entz et al. 1991; Lesser et al. 1996). Repair may be ongoing
(Lesser et al. 1994), may mostly occur after UV exposure
(A. T. Banaszak et al. unpubl. data), or may be largely absent
(Neale et al. 1998b). During exposure to polychromatic UV
1 PAR, a range of potential repair processes (Neale et al.
1994) is stimulated by longer wavelengths, thus counteract-
ing the damaging effects of UV radiation (Hirosawa and
Miyachi 1983; Samuelsson et al. 1985; Neale 1987). Low
temperatures have also been hypothesized to enhance the
effects of photoinhibition, most likely because of slower
rates of repair (Neale et al. 1994; Lesser et al. 1996). Low
nutrient levels may increase the photoinhibitory effect on
UV-exposed assemblages by limiting both photoprotective
and repair processes (Cullen and Lesser 1991; Behrenfeld et
al. 1994; Lesser et al. 1994; E. Litchman and P. J. Neale
unpubl. data). Thus, the relative importance of repair versus
protection will vary depending on specific conditions and
the physiological characteristics of the assemblage.
Most studies of the effects of UV radiation on phytoplank-
ton photosynthesis have concentrated on Antarctica, because
of the annually occurring, springtime ozone hole. Relatively
little is known about the responses of marine phytoplankton
in temperate regions with near-normal ozone levels, partic-
ularly over the full seasonal cycle. In particular, there has
594 Banaszak and Neale
Table 2. Notation
adet(l) m21 Detrital particulate (tripton) ab-
sorption coefficient
apart(l) m21 Total particulate absorption coeffi-
cient
a (l)*phy m2 mg Chl21 Phytoplankton absorbance normal-
ized to chlorophyll
E*inh Dimensionless Biologically effective fluence rate
for the inhibition of photosyn-
thesis
?(l) (mW m22)21 Biological weightings for inhibi-
tion of photosynthesis as a
function of UV irradiance
EPAR W m22 Photosynthetically available radia-
tion (400?700 nm)
Es W m22 Characteristic irradiance (EPAR) for
saturation of photosynthesis
l nm Wavelength
PB mg C mg Chl21
h21
Photosynthesis normalized to
chlorophyll a
PBs mg C mg Chl21
h21
Maximum rate of photosynthesis
in the absence of inhibition
UV, EUV mW m22 nm21 Ultraviolet radiation (280?400
nm)
Incident solar UV (290?400 nm)
UV-A mW m22 nm21 UV radiation between 320 and
400 nm
UV-B mW m22 nm21 UV radiation between 280 and
320 nm
been little study of shallow estuarine waters, the phytoplank-
ton dynamics of which are receiving increased attention be-
cause of the rising frequency of harmful algal blooms. Some
studies of other planktonic systems indicate that there is less
sensitivity to inhibition by UV radiation during the summer
months, compared with the spring or fall, but BWFs were
not determined (Hobson and Hartley 1983; Gala and Giesy
1991). Furgal and Smith (1997) measured inhibition in fresh-
water phytoplankton by natural solar UV-B and under con-
trolled conditions, using fluorescent lamps. Although solar
effects did vary throughout the year, UV inhibition under the
artificial irradiance was fairly constant.
The objective of this study was to determine the spectral
dependence of photoinhibition in estuarine phytoplankton
and to evaluate how the absolute sensitivities of phytoplank-
ton to UV radiation might change seasonally. Direct mea-
surements on natural phytoplankton populations are used to
assess the variability in sensitivity and to determine whether
a generalized BWF is suitable to characterize phytoplankton
photosynthetic response to UV radiation.
Materials and methods
Sample collection?Integrated water-column samples
were collected by use of a Labline 2L teflon sampler at
;1100 h on each of the dates indicated in Table 1. The
majority of the samples were collected from the Rhode Riv-
er, which is a turbid, eutrophic (Gallegos 1992) subestuary
on the western shore of the Chesapeake Bay in Maryland.
The primary sampling site (Sta. 4B) is located in an estuary
segment with a mean depth of 1.6 m (Jordan et al. 1991).
Two additional locations were sampled. One station (1A) is
located 1.4 km east of the mouth of the Rhode River, and
the other sample (Bay Bridge) was obtained on the eastern
shore of the bay close to the Chesapeake Bay Bridge. The
water samples were immediately transferred into an opaque
plastic bottle and maintained in the dark for ,1 h at ambient
water temperature until determination of photosynthesis-ir-
radiance (P-E) curves in the laboratory. The temperature
profile of the water column was measured at the time of
sampling by use of a Hydrolab temperature profiler, and Sec-
chi depth was measured with a 20-cm solid white disk. Ad-
ditional samples for optical measurements were taken in
1999. At these stations (Rhode River only), spectral irradi-
ance (2-nm bandwidth) was measured by use of a Satlantic
OCP 200 radiometer with filter center wavelengths of 325,
340, and 380 nm and 10 wavelengths (10-nm bandwidth) in
the visible range.
Photosynthesis measurements?Within 1 h of collection
of the water samples, photosynthesis as a function of UV
and PAR was measured by use of uptake of 14C-bicarbonate
and a modification of the ??photoinhibitron?? experimental
system, as described by Cullen and Neale (1997). An aliquot
was also used to simultaneously measure photosynthesis as
a function of PAR by use of a modification of the ??photo-
synthetron?? method (Lewis and Smith 1983) to obtain robust
estimates of P and Es (see Table 2 for notation). The pho-Bs
toinhibitron and the photosynthetron incubations were main-
tained at ambient water temperatures by use of recirculating
water baths. During the summer months, a 2.5-kW xenon
arc lamp was used as the light source for the photoinhibitron
experiments, whereas during other times of the year, a 1-kW
xenon arc lamp was sufficient to saturate photosynthesis.
The light-source beam was reflected upward by a mirror
through a recirculated heat trap onto quartz-bottom vials
containing the phytoplankton suspension inoculated with
14C-bicarbonate. The illuminated region was divided into
eight polychromatic sections by use of 5 3 5 cm Schott
series WG long-pass filters with nominal cutoffs (50% T) of
280, 295, 305, 320, 335, 345, and 360 nm and a Schott GG
400-nm long-pass filter as a control (essentially no UV ra-
diation). Each of the eight sections was further divided into
nine positions modified by insertion of neutral-density nickel
screens, resulting in 72 different combinations of spectral
composition and irradiance.
Irradiance measurements?Spectral irradiance [E(l), mW
m22 nm21] in each of the 72 treatments of the ??photoinhibi-
tron?? was measured by use of a calibrated 1,024-channel
diode-array spectroradiometer system (EG & G) consisting
of a quartz fiber-optic probe fitted with a diffuser appropriate
for measuring UV and PAR coupled to a spectrograph with
gratings suitable for UV and PAR regions (Acton Research)
(Cullen and Lesser 1991). The spectroradiometer was cali-
brated by scanning a mercury arc lamp for wavelength offset
and a 1-kW quartz halogen lamp (Eppley) traceable to the
National Institute of Standards and Technology for intensity.
For measuring UV (EUV), the lower-wavelength limit for sig-
nificant irradiance in the treatments was 282 nm, and the
595UV sensitivity of phytoplankton
upper-wavelength limit was 395 nm for the grating config-
uration. EPAR was calculated by use of the measured spectral
irradiance from 400 to 700 nm. Correction for stray light at
wavelengths ,310 nm was made by use of a Schott WG
345 nm long-pass filter (Cullen and Lesser 1991). Quanti-
tative characterization of the 72 spectral treatments was ob-
tained by use of principal component analysis (PCA) of
spectra of EUV normalized to EPAR, as described by Cullen
and Neale (1997). Spectral components common to all 72
treatments were calculated for each experiment. When PCA
was used, the large number of original wavelengths could
be adequately represented by two or three components (Nea-
le et al. 1994). Irradiance measurements in the ??photosyn-
thetron?? were made with a quantum scalar (4p) sensor
(Biospherical Instruments QSL-100).
Generation of biological weighting functions?Cullen et
al. (1992) developed an irradiance-dependent analytical
model (BWFE/P-I) in which photosynthesis is described as
a function of PAR and photoinhibition is dependent on bi-
ologically weighted UV irradiance and PAR (Cullen and
Neale 1997) such that
1
B BP 5 P tanh(E /E ) , (1)s PAR s
1 21 1 E*inh
where P is the maximum attainable rate of photosynthesisBs
in the absence of photoinhibition, EPAR is PAR expressed as
irradiance in energy units (W m22), and Es is a saturation
parameter for photosynthesis (c.f. Jassby and Platt 1976). PB
is the product of potential photosynthesis [P tanh(EPAR/Es)]Bs
and inhibition [1/(1 1 E )] and is the rate of photosynthesis*inh
normalized to chlorophyll (mg C mg Chl21 h21). The inhi-
bition term is a function of EUV, expressed as biologically
weighted exposure, E (dimensionless), where*inh
395 nm
E* 5 ?(l) 3 E(l) 3 Dl (2)
Oinh
l5282 nm
and ?(l) is the wavelength-dependent biological effective-
ness for inhibition of photosynthesis by UV (mW m22)21.
Although a parameter for EPAR-dependent inhibition (i.e.,
?PAR) can be included in Eq. 2, it was not needed. The BWF/
P-I model (Eq. 1) is similar to that described by Cullen and
Neale (1997), except that potential photosynthesis is de-
scribed by a hyperbolic tangent P-E function (Jassby and
Platt 1976), which gave higher R2 values for the model than
did fitting to an exponential function. The parameters of the
BWF/P-I model were determined statistically from results of
the 14C incorporation under a broad range of UVB : UVA :
PAR spectral ratios and absolute intensities of PAR, by use
of the PCA approach (Cullen et al. 1992; Neale et al. 1994;
Cullen and Neale 1997). Briefly, Eq. 2 is transformed to
express E in terms of spectral components, so that fitting*inh
the BWF requires estimation of only three or four coeffi-
cients (ho, the mean treatment effect over the whole irradi-
ance spectrum; hi, the component effect where i corresponds
to components 1, 2, or 3) to account for most of the variation
in PB. This method does not sacrifice spectral resolution and
requires no a priori assumptions about spectral slope, as is
the case for the Rundel approach (Rundel 1986). In most
cases, the best fit to the data required a two-component mod-
el, and, in rare instances, a three-component model was used.
No further significant increase in R2 was obtained by incor-
porating more components. The fitted model for the photo-
inhibitron measurements had an average R2 5 0.93 (mini-
mum 5 0.80, maximum 5 0.97). An average BWF was
computed for each season, with 95% confidence intervals
calculated from the standard errors of individual BWFs by
use of propagation of errors (Bevington 1969).
Independent measures of P and Es were also obtainedBs
from detailed (PAR only) P-E curves. Aliquots (1 ml) were
incubated in 7-ml scintillation vials by use of 38 different
light levels. Irradiance was provided by 250-W halogen
bulbs attenuated by use of neutral density screens and mea-
sured by use of a quantum scalar sensor (QSL-100) mounted
inside a scintillation vial. Quantum scalar irradiance (mmol
photons m22 s21) was converted to EPAR (W m22) by appli-
cation of a conversion factor (4.9 mmol J21) determined from
measured spectral irradiance. The photosynthetic parameters
(P and Es) were estimated as in Eq. 1 (i.e., with E 5 0).B *s inh
Curves were fitted with both the hyperbolic tangent function
and the exponential function. As for the photoinhibitron
data, the hyperbolic tangent function consistently gave better
fits than the exponential function.
Cellular absorption and chlorophyll concentrations?
Triplicate aliquots of phytoplankton samples were concen-
trated onto glass-fiber filters (Whatman GF/F) and scanned
against a wetted blank filter in a Cary IV spectrophotometer
from 280 to 750 nm, to estimate particulate-absorption co-
efficients. After scanning a fresh filter for total absorbance
(apart), the filters were then extracted into 100% methanol
overnight, rinsed, and rescanned, to determine detrital ab-
sorbance (adet). The phytoplankton absorption coefficient
(a ) was estimated from the difference apart 2 adet normal-*phy
ized to the chlorophyll concentration. Particulate absorbance
of samples collected in 1994 was estimated by illuminating
the filters with a 1-kW xenon arc source and measuring the
spectral transmission by use of the quartz fiber optic coupled
to a diode array spectrograph, as described earlier for UV
spectral characterization. Because we were interested mainly
in the correlation of absorbance and other optical parameters,
we did not correct for pathlength amplification (b). Chlo-
rophyll a concentrations in phytoplankton samples were de-
termined in triplicate aliquots, which were filtered onto
glass-fiber filters (Whatman GF/F), extracted into 10 ml of
90% acetone at 248C for 24 h, and measured fluorometri-
cally in a Turner Designs fluorometer. The fluorometer was
calibrated annually with spinach Chl a (Sigma).
MAA concentrations?The extraction and analysis of
MAAs by reverse-phase, isocratic, high-performance liquid
chromatography (HPLC) were performed according to the
procedures in Dunlap and Chalker (1986) with modifica-
tions. Approximately 50 ml of sample water was filtered
onto glass-fiber filters (Whatman GF/F), extracted overnight
into 100% HPLC-grade methanol at 48C, centrifuged, and
the supernatant used for MAA analysis. MAAs were sepa-
rated on a Brownlee RP-8 column (Spheri-5, 4.6 mm ID 3
250 mm), protected with an RP-8 guard column (Spheri-5,
596 Banaszak and Neale
Fig. 1. Seasonal variation in photosynthetic parameters of es-
tuarine phytoplankton from October 1994 to July 1996. Comparison
of (A) estimates of the chlorophyll-specific saturated rate of pho-
tosynthesis (P , mg C mg Chl21 h21); (B) estimates of the chloro-Bs
phyll-specific photosynthetic efficiency (a, mg C mg Chl21 h21
[mmol photons m22 s21]21); and (C) chlorophyll concentrations (mg
Chl m23) for each sample used in determining a biological weight-
ing function. Error bars show the standard error of the photosyn-
thetic parameter estimates. Shading corresponds to grouping of
dates into fall, winter, spring, or summer seasons.
4.6 mm ID 3 30 mm). The mobile phase, which consisted
of 25% methanol (v : v), 0.1% glacial acetic acid (v : v) in
water, was run at a flow rate of 0.7 ml min21. Identities of
peaks were resolved by the wavelength of maximal absor-
bance with confirmation by primary and secondary standards
by use of a diode array UV absorbance detector (Beckman
Gold System). All MAA concentrations were normalized to
Chl a, and concentrations are expressed in nmol (nmol chlo-
rophyll)21.
Incident irradiance?Total incident irradiance (cal cm22
d21) was measured with an Eppley pyroheliometer, and spec-
tral solar UV-B (minute averages) was measured over the
spectral range of 290?325 nm by use of a multifilter radi-
ometer with either 8 5-nm bandwidth filters (Correll et al.
1992) or 18 2-nm bandwidth filters (Early et al. 1998).
Results
A seasonal study of variation in sensitivity of Rhode River
phytoplankton to photoinhibition by UV radiation was begun
during October 1994 (Table 1). For the purposes of analysis,
seasons were divided into ??fall?? (23 September?22 Decem-
ber), ??winter?? (23 December?22 March), ??spring?? (23
March?22 June), and ??summer?? (23 June?22 September).
The number of replicates in each season was seven, except
in winter, when the number of replicates was two. The fall
and winter seasons tended to be diatom-dominated popula-
tions, whereas during spring, summer, and early fall, dino-
flagellates numerically dominated the water column. Other
photosynthetic flagellates were present year round in the wa-
ter samples but were usually less abundant than dinoflagel-
lates or diatoms (Table 1).
Photosynthetic parameters?The photosynthetic parame-
ters (Fig. 1) of the estuarine community varied seasonally.
The photosynthetron-based estimates of the light-saturated
rate of photosynthesis (P ) and the light saturation parameterBs
(Es) (n 5 38) had lower standard errors than fits of the same
parameters using the photoinhibitron. Although data from
the photoinhibitron has more degrees of freedom (n 5 72),
the design is not optimized for estimates of PAR-dependent
photosynthesis. The hyperbolic tangent?based P-E curves
were fitted with an R2 ranging from a minimum of 0.93 to
a maximum of 1.00, with an average of 0.98. Documentation
of these photosynthetic parameters allowed for an inventory
of the physiological state of the community at the time of
the determination of the BWF. No PAR photoinhibition was
observed up to 400 W m22 PAR. The chlorophyll-specific,
light-saturated rate of photosynthesis (P ) varied 20-foldBs
(Fig. 1A) over the seasons, whereas photosynthetic efficien-
cy (a 5 P /Es) varied 7-fold (Fig. 1B). Chlorophyll concen-Bs
trations varied 14-fold (Fig. 1C), with a general trend of low
chlorophyll concentrations during winter months and higher
chlorophyll concentrations during spring and summer
months.
Seasonal variation in biological weighting functions?P-
E curves generated by use of the photoinhibitron indicated
inhibition of photosynthesis, as determined by 14C uptake, at
saturating intensities when samples were exposed to UV ra-
diation. As progressively shorter wavelengths of UV-A then
UV-B were added to a constant background of PAR, pho-
toinhibition became progressively more severe (data not
shown, but see Neale et al. 1998a for an example data set).
An irradiance-dependent BWF/P-I model was chosen for
UV inhibition of Rhode River phytoplankton. When photo-
synthetic rates reach steady-state levels rapidly, photoinhi-
bition can be described as a function of irradiance (Neale
and Richerson 1987; Cullen and Lesser 1991). On the basis
of time courses of Rhode River phytoplankton photosynthe-
sis as inferred from pulsed amplitude?modulated fluorome-
try (data not shown), steady-state levels were reached within
;20 min of exposure to UV radiation.
When averaged over each season, BWFs for UV inhibi-
tion of photosynthesis in Chesapeake Bay phytoplankton do
not vary significantly between seasons (Fig. 2). Rhode River
assemblages are inhibited more strongly in the UV-B, par-
597UV sensitivity of phytoplankton
Fig. 2. BWFs for inhibition of photosynthesis by UV radiation
in phytoplankton from the Rhode River, Chesapeake Bay. Weight-
ing functions shown are the averages for fall 1994 and 1995 (n 5
7), winter 1995 (n 5 2), spring 1995 and 1996 (n 5 7), and summer
1995 and 1996 (n 5 7). Estimated 95% confidence intervals for
each average were calculated by propagation of errors from the
standard errors of individual BWFs. For clarity, only the 95% con-
fidence interval for the spring average is shown, the 95% confidence
intervals for the other averages are similar in width.
Fig. 3. Weighted solar irradiance determined by use of the prod-
uct of a BWF with solar irradiance. The BWF is the average (n 5
23) estimated for inhibition of photosynthesis by UV radiation for
phytoplankton from the Rhode River, Chesapeake Bay. The spectral
irradiance used for the weighting is a near-summer-solstice, noon-
time measurement (mW m22 nm21) taken at Table Mountain, Col-
orado (408N) with a total column ozone of 290 Dobson units (Early
et al. 1998) and smoothed to 1-nm effective bandwidth.
ticularly below 300 nm, but there is a significant influence
well into the UV-A. When the average BWF (n 5 23) for
Rhode River phytoplankton was used to weight a near sum-
mer solstice a spectral-irradiance curve recorded at a sim-
ilar latitude (Table Mountain, Colorado, 408N), effective
irradiance was highest in the range from 310 to 330 nm
(Fig. 3). There was considerable variation in the BWFs dur-
ing each season. Figure 4 shows the least sensitive (11 Jan-
uary 1995) and most sensitive (6 April 1995) BWFs ob-
tained for the Rhode River; these BWFs are significantly
different from each other (95% confidence intervals not
shown for clarity). There are differences in the sensitivity
or offset between the two BWFs, and there is variation in
the spectral shape or relative effects of UV-A versus UV-
B in these assemblages. Other BWFs for the Rhode River
are distributed fairly evenly between the minimum and
maximum BWFs, as shown by the variation in weightings
at 310 nm [?(310)] as a function of time (Fig. 5A). The
photoprotective properties of the phytoplankton also do not
show any seasonal trends. Figure 5B shows the variation
in MAA content that ranged from undetectable to ;6 nmol
(nmol Chl)21. The mean ratio of phytoplankton particulate
absorbance (a ) at 310?665 nm is 1.5 (data not shown),*phy
consistent with the presence of moderate concentrations of
UV-absorbing compounds.
Environmental variation?Although there was little sea-
sonal variation in ?(310), there was seasonal variation in
several other basic environmental parameters. Temperature
is highly variable in this shallow-water environment, ranging
from ,58C in the winter to .308C in the summer (Fig. 5C).
A general indication of water transparency was provided by
Secchi depth, which was ;0.5 m during the late spring and
summer but was usually .1.0 m during the fall and winter.
To better understand spectrally specific changes in water
transparency, we examined the relationship between partic-
ulate absorbance and attenuation coefficient. This could be
done using a more recent (1999) data set of particulate ab-
sorbance (measured as described) and spectral attenuation
(Kd) at 325 and 340 nm and for PAR, measured in the Rhode
River by use of a Satlantic OCP 200 profiling spectroradi-
ometer (C. L. Gallegos et al. unpubl. data). In this data set
(not shown), we focused on 340 nm as a general indicator
of UV transparency. Both total [apart(340)] and detrital
[adet(340)] particulate absorbance are correlated with Kd at
340 nm, but the strongest association is with detrital absor-
bance (R2 5 0.67). This indicates that detritus and detrital
correlates (e.g., dissolved organic matter) are the dominant
factors influencing variations in UV attenuation of these
shallow (1.5 m average in the Rhode River) waters (Kirk
1994).
The adet(340) of samples used for the BWF measurements
suggest a general seasonal cycle of UV water transparency
with lowest attenuation (highest transparency) in the winter
and greatest attenuation in the summer (Fig. 5D), consistent
with the Secchi depths. The sample from the deeper-water
(7 m) Bay Bridge station (12 July 1995) is an exception to
this pattern. Here, adet(340) was comparatively low (0.8 m21),
598 Banaszak and Neale
Fig. 4. Comparison of absolute spectral weights for UV inhi-
bition of photosynthesis in natural phytoplankton assemblages from
the Rhode River with high (spring) and low (winter) sensitivity with
published irradiance-dependent BWFs. The published BWFs are for
a natural phytoplankton assemblage from Lake Bonney, Antarctica
(Neale et al. 1994); a laboratory culture of the dinoflagellate Pro-
rocentrum micans (Cullen et al. 1992); a laboratory culture of the
diatom Phaeodactylum sp. (Cullen et al. 1992); low- and high-light-
acclimated laboratory cultures of the dinoflagellate, G. sanguineum
(Neale et al. 1998a), and a natural diatom-dominated assemblage
from McMurdo Sound, Antarctica maintained in outdoor culture
(Neale et al. 1994).
Fig. 5. Seasonal variation in (A) spectral weighting function for
inhibition of photosynthesis at 310 nm [?(310), (mW m22)21], (B)
total concentration of MAAs (nmol [nmol Chl]21), (C) water tem-
perature (8C), and (D) detrital particulate absorbance (adet(340),
m21). The error bars show the standard error (n 5 3) of MAA
concentrations and the 95% confidence intervals of ?[310].
even though the Secchi depth was shallow (0.75 m). This
suggests that our empirical relationship between adet(340)
and Kd only holds in shallow waters. The seasonal cycle of
water transparency inferred from adet(340) is the reverse of
the seasonal variation in solar irradiance, particularly in the
UV. The seasonal variation of UV in the water column is
less than the seasonal variation in incident irradiance (Table
3). A similar analysis based on the correlation of adet(340)
and Kd for PAR in the Rhode River indicates that changes
in water transparency had less impact on the seasonal vari-
ation of PAR (Table 3). Seasonal variation of nutrient avail-
ability is well established for the Rhode River (Gallegos et
al. 1992; Jordan et al. 1992). In particular, nitrate is usually
abundant in the late fall through early spring but low the
rest of the year (except for episodic watershed events). Nu-
trients were not measured for the samples used for BWF
analysis, but routine measurements of the Rhode River dur-
ing the period indicate that nitrate was low (,1 mM) during
the whole sample period, except for the winter 1995 and
early spring 1996 (27 March) samples (T. Jordan pers.
comm.).
Impact of UV on primary production?The main objective
of this study was to understand the seasonal variation of
BWFs, and the optical data necessary to make a compre-
hensive assessment of in situ effects were not measured.
Nevertheless, a first-order sensitivity analysis was performed
to ascertain the potential for UV impacts on Rhode River
primary production. The details of this calculation can be
obtained from P. Neale. The sensitivity analysis used obser-
vations of surface spectral irradiance (noon, 4-week period
around summer solstice 1999; c.f. Fig. 3), optical penetration
(weekly profiles summer 1999), and BWFs (this study) to
599UV sensitivity of phytoplankton
Table 3. Seasonal variation of incident and in situ irradiance in
the Rhode River. Irradiance was averaged for the 1 week period
prior to sampling (Table 1), from which a seasonal average was
calculated grouping samples as in Fig. 5. Average irradiance over
the water column was estimated by use of the formula E0(1 2 e2kz)/
kz, where E0 is incident irradiance, k is the attenuation coefficient,
and depth (z) is 1.5 m. Integrated daily irradiance (cal cm22 d21)
measured with an Eppley pyroheliometer was converted to mol pho-
tons m22 d21 of PAR by use of the factor 9.6 3 1026 mol photons
cal21. The attenuation coefficient for PAR (KdPAR) was estimated
from adet (340) on the basis of regression analysis of 1999 data (R2
5 0.47). Incident UV-B for the 1-h period centered around solar
noon in the 320?325-nm wavelength range (mW m22) was mea-
sured with a Smithsonian multifilter (SR8 or SR18) spectroradi-
ometer. The attenuation coefficient for 325 nm [(Kd(325)] was es-
timated by application of the average ratio of Kd(325) : Kd(340)(1.3)
to K(340), estimated from regression with adef(340) in 1999 data (R2
5 0.67). The last column is the ratio between the maximum and
minimum value in each row.
Winter Spring Summer Fall
Max/
Min
Incident
PAR (mol m22 d21)
UV (320?325 nm)
(mW m22)
21.5
401.0
33.9
877.6
42.2
879.1
21.3
612.1
2.0
2.2
In situ
PAR (mol m22 d21)
UV (320?325 nm)
(mW m22)
8.0
39.0
11.6
70.3
13.2
62.8
6.8
44.4
2.0
1.6
calculate depth-integrated primary production to 1.6 m. Ac-
cording to this analysis, midday primary production in the
presence of UV is 3%?33% lower than PAR-only produc-
tion, depending on conditions. Variation in sensitivity and
optical properties (UV and PAR transparency) contributed
about equally to the variation in inhibition by UV. Optical
variation may be even more important on short timescales,
because time series (1-h resolution) of optical properties in
the Rhode River have revealed many-fold changes in trans-
parency over the course of a few hours (Gallegos and Neale;
unpublished data).
Discussion
We have determined 23 BWFs for inhibition of phyto-
plankton photosynthesis, which constitute the first annual
survey of BWFs for phytoplankton (estuarine or otherwise)
in marine waters. Phytoplankton assemblages in the Rhode
River were sensitive to UV radiation throughout the year,
and both UV-A and UV-B contributed to inhibition, at least
by surface irradiance. There was no significant inhibition of
photosynthesis by PAR. There was little variation in seasonal
average BWFs through the year, but individual BWFs were
quite variable. Between the least and most sensitive BWFs
for the Rhode River, there are differences in both the abso-
lute magnitude and in the relative effect of UV-B versus UV-
A (Fig. 4). The weighting at 310 nm varied by approximately
one order of magnitude (Fig. 5A).
Comparison with other biological weighting functions?
The range of variation observed in the Rhode River is com-
parable to the overall variation in other irradiance-dependent
(E model) BWFs for natural assemblages and cultures (Fig.
4). This variation is seen both in marked differences in the
offset or magnitude of the BWF over all wavelengths as well
as differences in the shapes of the BWFs. The highest sen-
sitivity BWF observed to date (using the E model) is for a
cryptomonad-dominated assemblage from the extreme shade
environment of a permanently ice-covered lake in Antarctica
(Lake Bonney; Neale et al. 1994). The BWF of the most-
sensitive spring assemblage from the Rhode River was com-
parable to the Lake Bonney BWF in the UV-B and short-
wavelength UV-A. A lower sensitivity group consists of
laboratory cultures of diatoms and dinoflagellates (Cullen et
al. 1992; Neale et al. 1998a), and a natural, diatom-domi-
nated assemblage from McMurdo Sound, Antarctica main-
tained in outdoor culture (Neale et al. 1994). The BWF of
the low sensitivity, winter assemblage of phytoplankton from
the Rhode River was comparable to this group.
Other Rhode River BWFs were about evenly distributed
between these high- and low-sensitivity extremes (c.f. Fig.
5A), so the average BWF for the Rhode River corresponds
to a moderate level of sensitivity. A large number of BWFs
have also been determined for Antarctic assemblages, in-
cluding the McMurdo area (Neale et al. 1994), the Weddell-
Scotia Confluence (WSC; Neale et al. 1998b), and coastal
waters of the Antarctic Peninsula (Neale et al. in press; J. J.
Fritz et al. unpubl. data). We sought to determine how the
average BWF for these environments resembled that of the
Rhode River. However, the Antarctic BWFs include both
irradiance (E model) and cumulative exposure (H model)
BWFs; the latter were estimated for WSC phytoplankton
(Neale et al. 1998b; Neale 2000). We established a common
basis for comparing BWFs with these two models by lim-
iting the comparison to effects over the 1-h period used to
measure photosynthesis. This procedure requires adjustment
for both the units and functional differences between the H
and E models in describing the hyperbolic dependence of
inhibition on exposure (Neale et al. in press). Adjusted in
this way, average BWFs for the Rhode River and the Ant-
arctic assemblages are quite similar (Fig. 6). Indeed, the sim-
ilarity is quite striking, considering how different are these
two environments. In each case, the average response to UV
shows significantly greater sensitivity than is typically
shown by laboratory cultures, particularly in the UV-A (c.f.
Fig. 4) (Neale and Kieber 2000). It appears that natural phy-
toplankton are typically constrained in their ability to defend
against UV, although more fully acclimated assemblages
sometimes occur (e.g., low-sensitivity Rhode River and Mc-
Murdo culture BWFs in Fig. 4). The possible causes of BWF
variation are considered below in the section, Seasonal var-
iability of UV response.
Several other studies of temperate and tropical marine
phytoplankton have shown sensitivity of photosynthesis to
UV but have not defined BWFs (e.g., Smith et al. 1980;
Helbling et al. 1993). The only study that defines a BWF
for non-Antarctic marine phytoplankton is that of Behrenfeld
et al. (1993). These authors suggested a single BWF was
effective in estimating inhibition of photosynthesis by UV-
600 Banaszak and Neale
Fig. 6. Average BWFs for the inhibition of photosynthesis by
UV for Rhode River and Antarctic phytoplankton (Neale and Kieber
2000) and the generalized BWF defined by Behrenfeld et al. (1993).
The Behrenfeld et al. BWF is given as inhibition per unit irradiance
in mW m22, under the assumption of a 1-h exposure. It was cal-
culated as the product of the relative action spectrum (unity at 300
nm), the response coefficient for inhibition by photosynthesis
(0.0116 reciprocal [J m22]) and a constant (3.6) to convert effect
per J to effect per mW-h.
Table 4. Irradiance-dependent biological weighting functions (BWFs) published to date and the mycosporine-like amino acid (MAA)
concentrations of the population used for the determination of the BWF. The high- and low-sensitivity BWFs are diagrammed in Fig. 4.
Assemblage Reference Culture and/or light conditions
MAA [nmol
(nmol Chl a)21]
High-Sensitivity BWFs
Lake Bonney cryptomonad assemblage
Rhode River spring assemblage
Neale et al. (1994)
This study
Variable but low light
Variable
0?
0.05 (n 5 1)
Low-Sensitivity BWFs
Rhode River winter assemblage
Gymnodinium sanguineum low-light culture
This study
Neale et al. (1998a)
Variable
Static 15 W m22
1.66 (n 5 1)
2.09 (n 5 5)
G. sanguineum high-light culture
McMurdo Sound diatoms (outdoor culture)
Neale et al. (1998a)
Neale et al. (1994)
Static 76 W m22
Variable noon 5 125 W m22
average 5 37 W m22
44.35 (n 5 5)
3.91 (n 5 6)
B over a range of marine systems (although they did not
sample estuaries). A general comparison of the Behrenfeld
et al. BWF with Rhode River BWFs is not possible, because
this previous study models inhibition as a linear function of
cumulative exposure, whereas the BWF/P-I model assumes
a hyperbolic dependence on weighted irradiance. A specific
comparison can be made over a 1-h incubation period, just
as for the WSC BWFs, although we cannot adjust for the
linear versus hyperbolic dependence on exposure. If the Beh-
renfeld et al. BWF is multiplied by 3.6 (to account for the
number of Joules in a mW-h), the weight at 300 nm is similar
to the ?(300) of the Rhode River and Antarctic BWFs (Fig.
6). However, average weights for the latter BWFs are one
to two orders of magnitude higher than the Behrenfeld et al.
BWF in the long-wavelength UV-B and short-wavelength
UV-A range, where solar radiation has maximum biological
effectiveness (c.f. Fig. 3). Because the WSC BWFs were
based on a cumulative exposure model (H model), they can
be directly compared with the Behrenfeld et al. BWF (Neale
et al. 1998b). However, a similar result is obtained: the
weights are similar at 300 nm and diverge at longer wave-
lengths (Neale 2000). On the other hand, there does seem to
be some constancy in responses to UV by natural phyto-
plankton assemblages when considering averages over large
space- and timescales (Fig. 6), which recalls Behrenfeld et
al.?s (1993) conjecture. However, the average BWFs pre-
sented here differ considerably from the hypothesized com-
mon response. At this point, it is unknown whether the av-
erage response in other marine systems will resemble the
Rhode River and Antarctic averages. Finally, it is important
to note that these average BWFs only apply to a 1-h time-
scale, and the response of a specific phytoplankton assem-
blage can diverge quite substantially from these averages.
MAAs and UV sensitivity?Accumulation of MAAs has
received much attention as a primary defense against UV in
marine organisms. Higher concentrations of MAAs have
been correlated with UV photoprotection in dinoflagellates
(Banaszak and Trench 1995; Neale et al. 1998a). High con-
centrations of MAAs can decrease sensitivity to UV, as has
been shown for G. sanguineum, a dinoflagellate of moderate
cell size. The shape of the BWF was affected when total
MAA content was ;44 nmol (nmol Chl)21 but not when
MAA content was 2 nmol (nmol Chl)21 (Neale et al. 1998a).
In the Rhode River, the average MAA concentration was 1.5
nmol (nmol Chl)21. This suggests that the protective effects
of MAAs on photosynthesis in the Rhode River were usually
too small to resolve a spectral screening. For two samples
from the Rhode River, MAA content was higher [between 5
and 6 nmol (nmol Chl)21]. MAAs may have provided some
photoprotection on these dates, but the effect still may have
been too small to resolve without replicate BWF determi-
nations, as was even necessary at 44 nmol (nmol Chl)21. On
601UV sensitivity of phytoplankton
the other hand, assemblages with low concentrations of
MAAs are characterized by high sensitivity to UV radiation
(Table 4). Lake Bonney cryptomonads were not examined
for MAAs; however, Vernet et al. (1994) showed that cryp-
tomonads from Antarctica have a low spectral-absorption co-
efficient at 300 nm compared with PAR, suggesting that
these cryptomonads have either no or very low concentra-
tions of MAAs. For G. sanguineum, the ?(310) was low even
when MAA concentration was comparable to the average
concentration in the Rhode River assemblages. Even though
MAA-accumulating phytoplankton are generally more resis-
tant to UV, the MAAs themselves may only be a second
level of defense (Lesser 1996a,b) that affect the spectral
shape of the BWF more than the absolute sensitivity, such
as was seen in high-light? versus low-light?grown G. san-
guineum (Neale et al. 1998a).
Seasonal variability of UV response?Seasonal variation
of environmental parameters is a basic characteristic of most
temperate aquatic ecosystems. The Rhode River environ-
ment is seasonally variable in temperature, as well as in light
and nutrient availability (Fig. 5, Table 3). Rhode River phy-
toplankton respond to this seasonal variation in several ways.
Photosynthetic performance (i.e., P ) is highest during theBs
summer, and phytoplankton biomass is high during spring
and summer (Fig. 5). Summer blooms are usually dominated
by dinoflagellates, whereas other flagellates are more im-
portant in the winter and spring, and diatoms are found year-
round (C. Gallegos pers. comm.). The UV sensitivity of phy-
toplankton has been shown to differ depending on light
acclimation status (Neale et al. 1998a), nutrient availability
(Cullen and Lesser 1991; Behrenfeld et al. 1994; Lesser et
al. 1994), and temperature (Roos and Vincent 1998), and
there are some systematic differences between sensitivity of
algal groups, or at least species (Xiong et al. 1996; E. Litch-
man and P. J. Neale unpubl. data). As a result, seasonal shifts
in UV sensitivity could be expected, and simple measures
of UV sensitivity have been seasonally variable in some sys-
tems (Hobson and Hartley 1983; Gala and Giesy 1991).
Despite these expectations, seasonally averaged BWFs for
inhibition of photosynthesis did not vary in any systematic
way. This was not due to an absence of variability per se
but a result of the temporal distribution of variability. For
?(310), the coefficient of variation was 50% over all sam-
ples; however, only 8% of variance occurred between sea-
sons (1-way ANOVA). The remaining variance (92%) oc-
curred within seasons (i.e., between samples). Presently, we
cannot account for this surprising temporal distribution. It
may be that factors that contributed to short-term variability
tended to covary in such a way as to minimize shifts in UV
response on a seasonal basis. Multiple regression analysis of
?(310) on temperature, irradiance, and MAAs were attempt-
ed to test this hypothesis empirically, but no significant (P
, 0.05) relationships could be derived. Nutrients were not
included in this analysis because they were not measured for
the BWF samples. This suggests that our database is pres-
ently too small to resolve relationships and/or that variations
were caused by other factors that were not measured (e.g.,
nutrient availability).
We do have some indications from other studies as to how
species composition, light, temperature, and nutrient avail-
ability can affect sensitivity to UV, but our understanding is
not presently advanced enough to formulate a quantitative
model of BWF response to multiple factors. Nevertheless, it
is instructive to consider conceptually how several factors or
factor interactions may have contributed to the variation that
we have documented for the Rhode River assemblages. Of
the factors listed, probably the least is known about variation
in UV sensitivity between species, a detailed analysis of
which has not been attempted for our data set. However, we
did observe that the samples with the highest sensitivity (ear-
ly spring 1995) were the ones with the highest proportion
of flagellates such as Cryptomonas sp., whereas the low-
sensitivity (winter 1995) sample was dominated by a dino-
flagellate (Katodinium rotundatum). The comparative sen-
sitivity of dinoflagellates and cryptomonads is being further
investigated in current culture studies. The availability of
PAR and water temperature can influence UV sensitivity be-
cause they determine the light-acclimation status of phyto-
plankton (MacIntyre et al. 2000). High light acclimation in
algae is correlated with increased antioxidants such as ca-
rotenoids and with increased capacity for turnover of dam-
aged molecular complexes (??repair??). Antioxidants can con-
tribute to UV resistance to the extent that reactive oxygen
species mediate UV effects (reviewed in Vincent and Neale
2000; Neale and Kieber 2000). Repair processes also lower
biological weights by affecting overall sensitivity and pos-
sibly by altering the shape of the BWF, depending on the
wavelength specificity of the damage and repair processes.
The twofold seasonal increase in available irradiance (Ta-
ble 3) might seem sufficient to induce a higher degree of
light acclimation in summer assemblages, but there is a com-
plication. This is because acclimation is a function of both
light availability, which controls the generation of energy
carriers (NADPH and ATP), and other factors, like temper-
ature, that determine metabolic rates and thus energy con-
sumption?a concept called excitation pressure (Maxwell et
al. 1995) or energy balance (MacIntyre et al. 2000). Maxwell
et al. (1995) found that the light intensity needed to induce
resistance to inhibition by EPAR was higher for Chlorella vul-
garis grown at high temperatures than that for cultures
grown at low temperatures. In the Rhode River, both irra-
diance and temperature increase seasonally, but the relative
increase in temperature is larger (Fig. 5B, Table 3). It may
be that excitation pressure is at a similar level, on average,
between seasons but is variable within seasons. However, a
more detailed data set is needed to test this hypothesis. Pre-
vious studies in the Rhode River have documented brief (i.e.,
several day) phytoplankton blooms, which are induced by
episodic nutrient enrichment (Gallegos 1992; Gallegos et al.
1992), and short-term forcing of photosynthesis is probably
a general feature of shallow estuarine environments (Mac-
Intyre et al. 2000).
Another factor that can constrain acclimation to UV ex-
posure is low nutrient availability. Nutrient stress, particu-
larly limited availability of nitrogen, may affect photopro-
tection because of the requirement of nitrogen for MAA
synthesis. Preliminary results indicate that accumulation of
MAAs is decreased under conditions of nitrogen limitation
(E. Litchman et al. unpubl. data). Rates of repair are also
602 Banaszak and Neale
slower under nitrogen-limited conditions, probably because
nitrogen-requiring enzymes are needed for repair (Cullen
and Lesser 1991; E. Litchman unpubl. data). Seasonal nitro-
gen availability is usually lowest when irradiance is highest.
Again, there is a possibility that the opposed seasonal phas-
ing of environmental factors may limit seasonal variation in
BWFs. Because average responses in the Rhode River phy-
toplankton resemble those of the Antarctic, and perhaps oth-
er systems, a better understanding of what causes the vari-
ability in this system may have implications beyond the
estimation of effects in shallow estuarine waters.
In summary, we have shown that Rhode River phyto-
plankton have a moderate sensitivity to UV that mainly
varies on short-term rather than seasonal timescales. The
BWFs provide an improved basis for assessing the effects
of long-term changes in the UV environment on the phyto-
plankton community, as can occur through changes in inci-
dent UV and UV transparency of the water body. Better
understanding of the effects of UV variation on estuarine
productivity and species composition is dependent on ob-
taining more information on short-term variation in estuarine
optical properties and the kinetics of acclimation to UV ra-
diation by phytoplankton.
References
BANASZAK, A. T., AND R. K. TRENCH. 1995. Effects of ultraviolet
(UV) radiation on marine microalgal-invertebrate symbioses.
II. The synthesis of mycosporine-like amino acids in response
to exposure to UV in Anthopleura elegantissima and Cassio-
peia xamachana. J. Exp. Mar. Biol. Ecol. 194: 233?250.
BEHRENFELD, M. J., J. W. CHAPMAN, J. T. HARDY, AND H. LEE II.
1993. Is there a common response to ultraviolet-B radiation by
marine phytoplankton? Mar. Ecol. Prog. Ser. 102: 59?68.
, H. I. LEE, AND L. F. SMALL. 1994. Interactions between
nutritional status and long-term responses to ultraviolet-B ra-
diation stress in a marine diatom. Mar. Biol. 118: 523?530.
BEVINGTON, P. R. 1969. Data reduction and error analysis for the
physical sciences. McGraw Hill.
BOTHWELL, M. L., D. M. J. SHERBOT, AND C. M. POLLOCK. 1994.
Ecosystem response to solar ultraviolet-B radiation: Influence
of trophic-level interactions. Science 265: 97?100.
CORRELL, D. L., C. O. CLARK, B. GOLDBERG, V. R. GOODRICH, D.
R. HAYES, JR., W. H. KLEIN, AND W. D. SCHECHER. 1992.
Spectral ultraviolet-B radiation fluxes at the Earth?s surface:
Long term variations at 398N, 778W. J. Geophys. Res. 97(D7):
7579?7591.
CULLEN, J. J., AND M. P. LESSER. 1991. Inhibition of photosynthesis
by ultraviolet radiation as a function of dose and dosage rate:
Results for a marine diatom. Mar. Biol. 111: 183?190.
, AND P. J. NEALE. 1997. Biological weighting functions for
describing the effects of ultraviolet radiation on aquatic sys-
tems, p. 97?118. In D.-P. Ha?der [ed.], Effects of ozone deple-
tion on aquatic ecosystems. R. G. Landes.
, , AND M. P. LESSER. 1992. Biological weighting
function for the inhibition of phytoplankton photosynthesis by
ultraviolet radiation. Science 258: 646?650.
DEMERS, S., S. ROY, R. GAGNO, AND C. VIGNAULT. 1991. Rapid
light-induced changes in cell fluorescence and in xanthophyll-
cycle pigments of Alexandrium excavatum (Dinophyceae) and
Thalassiosira pseudonana (Bacillariophyceae): A photoprotec-
tion mechanism. Mar. Ecol. Prog. Ser. 76: 185?193.
DUNLAP, W. C., AND B. E. CHALKER. 1986. Identification and quan-
titation of near-UV absorbing compounds (S-320) in a her-
matypic scleractinian. Coral Reefs 5: 155?159.
EARLY, E., AND OTHERS. 1998. The 1995 North American inter-
agency intercomparison of ultraviolet monitoring spectroradi-
ometers. J. Res. Natl. Inst. Stand. Technol. 103: 15?62.
FURGAL, J. A., AND R. E. H. SMITH. 1997. Ultraviolet radiation and
photosynthesis by Georgian Bay phytoplankton of varying nu-
trient and photoadaptive status. Can. J. Fish. Aquat. Sci. 54:
1659?1667.
GALA, W. R., AND J. P. GIESY. 1991. Effects of ultraviolet radiation
on the primary productivity of natural phytoplankton assem-
blages in Lake Michigan. Ecotoxicol. Environ. Saf. 22: 345?
361.
GALLEGOS, C. L. 1992. Phytoplankton photosynthesis, productivity,
and species composition in a eutrophic estuary: Comparison of
bloom and non-bloom assemblages. Mar. Ecol. Prog. Ser. 81:
257?267.
, T. E. JORDAN, AND D. L. CORRELL. 1992. Event-scale re-
sponse of phytoplankton to watershed inputs in a subestuary:
Timing, magnitude, and location of blooms. Limnol. Oceanogr.
37: 813?828.
HELBLING, E. W., AND OTHERS. 1993. Marine phytoplankton re-
sponse to ultraviolet radiation in middle latitudes (33 degree
S). Rev. Biol. Mar. 28: 219?237.
, V. VILLAFAN? E, M. FERRARIO, AND O. HOLM-HANSEN.
1992. Impact of natural ultraviolet radiation on rates of pho-
tosynthesis and on specific marine phytoplankton species. Mar.
Ecol. Prog. Ser. 80: 89?100.
HIROSAWA, T., AND S. MIYACHI. 1983. Inactivation of Hill reaction
by long-wavelength radiation (UV-A) and its photoreactivation
by visible light in the cyanobacterium, Anacystis nidulans.
Arch. Microbiol. 135: 98?102.
HOBSON, L. A., AND F. HARTLEY. 1983. Ultraviolet irradiance and
primary production in a Vancouver Island fjord, British Colum-
bia, Canada. J. Plankton Res. 5: 325?331.
JASSBY, A. D., AND T. PLATT. 1976. Mathematical formulation of
the relationship between photosynthesis and light for phyto-
plankton. Limnol. Oceanogr. 21: 540?547.
JORDAN, T. E., D. L. CORRELL, J. MIKLAS, AND D. E. WELLER.
1991. Nutrients and chlorophyll at the interface of a watershed
and an estuary. Limnol. Oceanogr. 36: 251?267.
KARENTZ, D., J. E. CLEAVER, AND D. L. MITCHELL. 1991. Cell
survival characteristics and molecular responses of Antarctic
phytoplankton to ultraviolet-B radiation. J. Phycol. 27: 326?
341.
KIRK, J. T. O. 1994. Light and photosynthesis in aquatic ecosystems,
2nd ed. Cambridge Univ. Press.
LESSER, M. P. 1996a. Acclimation of phytoplankton to UV-B radi-
ation: Oxidative stress and photoinhibition of photosynthesis
are not prevented by UV-absorbing compounds in the dinofla-
gellate Prorocentrum micans. Mar. Ecol. Prog. Ser. 132: 287?
297
. 1996b. Erratum. Mar. Ecol. Prog. Ser. 141: 312.
, J. J. CULLEN, AND P. J. NEALE. 1994. Carbon uptake in a
marine diatom during acute exposure to ultraviolet B radiation:
Relative importance of damage and repair. J. Phycol. 30: 183?
192.
, P. J. NEALE, AND J. J. CULLEN. 1996. Acclimation of Ant-
arctic phytoplankton to ultraviolet radiation: Ultraviolet-ab-
sorbing compounds and carbon fixation. Mol. Mar. Biol. Bio-
technol. 5: 314?325.
LEWIS, M. R., AND J. C. SMITH. 1983. A small volume, short-in-
cubation-time method for measurement of photosynthesis as a
function of incident irradiance. Mar. Ecol. Prog. Ser. 13: 99?
102.
LUBIN, D., B. G. MITCHELL, J. E. FREDERICK, A. D. ALBERTS, C.
603UV sensitivity of phytoplankton
R. BOOTH, T. LUCAS, AND D. NEUSCHULER. 1992. A contri-
bution towards understanding the biospherical significance of
Antarctic ozone depletion. J. Geophys. Res. 97: 7817?7828.
MACINTYRE, H. L., T. M. KANA, AND R. J. GEIDER. 2000. The effect
of water motion on short-term rates of photosynthesis by ma-
rine phytoplankton. Trends Plant Sci. 5: 12?17.
MAXWELL, D. P., S. FALK, AND N. P. A. HUNER. 1995. Photosystem
II excitation pressure and development of resistance to photo-
inhibition. I. Light-harvesting complex II abundance and zea-
xanthin content in Chlorella vulgaris. Plant Physiol. 107: 687?
694.
NEALE, P. J. 1987. Algal photoinhibition and photosynthesis in the
aquatic environment, p. 39?65. In D. J. Kyle, C. B. Osmond,
and C. J. Arntzen [eds.], Photoinhibition. Elsevier.
. 2000. Spectral weighting functions for quantifying the ef-
fects of ultraviolet radiation in marine ecosystems, p. 73?100.
In S. J. de Mora, S. Demers, and M. Vernet [eds.], The effects
of UV radiation on marine ecosystems. Cambridge Univ. Press.
, A. T. BANASZAK, AND C. R. JARRIEL. 1998a. Ultraviolet
sunscreens in Gymnodinium sanguineum (Dinophyceae): My-
cosporine-like amino acids protect against inhibition of pho-
tosynthesis. J. Phycol. 34: 928?938.
, J. J. CULLEN, AND R. F. DAVIS. 1998b. Inhibition of marine
photosynthesis by ultraviolet radiation: Variable sensitivity of
phytoplankton in the Weddell-Scotia Confluence during the
austral spring. Limnol. Oceanogr. 43: 433?448.
, R. F. DAVIS, AND J. J. CULLEN. 1998c. Interactive effects
of ozone depletion and vertical mixing on photosynthesis of
Antarctic phytoplankton. Nature 392: 585?589.
, J. J. FRITZ, AND R. F. DAVIS. In press. Effects of UV on
photosynthesis of Antarctic phytoplankton: Models and appli-
cation to coastal and pelagic assemblages. Rev. Chil. Hist. Nat.
, AND D. J. KIEBER. 2000. Assessing biological and chemical
effects of UV in the marine environment: Spectral weighting
functions, p. 61?83. In R. E. Hester and R. M. Harrison [eds.],
Causes and environmental implications of increased U.V.-B.
radiation. Issues in Environmental Science and Technology, 14,
Royal Society of Chemistry.
, M. P. LESSER, AND J. J. CULLEN. 1994. Effects of ultraviolet
radiation on the photosynthesis of phytoplankton in the vicinity
of McMurdo Station. Antarct. Res. Ser. 62: 125?142.
, AND P. J. RICHERSON. 1987. Photoinhibition and the diurnal
variation of phytoplankton photosynthesis. I. Development of
a photosynthesis-irradiance model from studies of in situ re-
sponses. J. Plankton Res. 9: 167?193.
RUNDEL, R. D. 1986. Action spectra and estimation of biologically
effective UV radiation. Physiol. Plant. 58: 360?366.
ROOS, J. C., AND W. F. VINCENT. 1998. Temperature dependence of
UV radiation effects on Antarctic cyanobacteria. J. Phycol. 34:
118?125.
SAMUELSSON, G., A. LO? NNEBORG, E. ROSENQVIST, P. GUSTAFSON,
AND G. O? QUIST. 1985. Photoinhibition and reactivation of pho-
tosynthesis in the cyanobacterium Anacystis nidulans. Plant
Physiol. 79: 992?995.
SMITH, R. C., K. S. BAKER, O. HOLM-HANSEN, AND R. S. OLSON.
1980. Photoinhibition of photosynthesis in natural waters. Pho-
tochem. Photobiol. 31: 585?592.
, AND OTHERS. 1992. Ozone depletion: Ultraviolet radiation
and phytoplankton biology in antarctic waters. Science 255:
952?959.
VERNET, M., E. A. BRODY, O. HOLM-HANSEN, AND B. G. MITCH-
ELL. 1994. The response of antarctic phytoplankton to ultra-
violet radiation: Absorption, photosynthesis, and taxonomic
composition. Antarct. Res. Ser. 62: 143?158.
VINCENT, W. F., AND P. J. NEALE. 2000. Mechanisms of UV damage
in aquatic organisms, p. 149?176. In S. J. de Mora, S. Demers,
and M. Vernet [eds.], The effects of UV radiation on marine
ecosystems. Cambridge Univ. Press.
XIONG, F. S., F. LEDERER, J. LUKAVSKY, AND L. NEDBAL. 1996.
Screening of freshwater algae (Chlorophyta, Chromophyta) for
ultraviolet-B sensitivity of the photosynthetic apparatus. J.
Plant Physiol. 148: 42?48.
Received: 10 May 2000
Accepted: 14 October 2000
Amended: 18 December 2000