1843
Ecological Applications, 11(6), 2001, pp. 1843?1857
q 2001 by the Ecological Society of America
BENEFICIAL AND DETRIMENTAL EFFECTS OF UV ON AQUATIC
ORGANISMS: IMPLICATIONS OF SPECTRAL VARIATION
CRAIG E. WILLIAMSON,1,3 PATRICK J. NEALE,2 GABRIELLA GRAD,1 HENDRIKA J. DE LANGE,1
AND BRUCE R. HARGREAVES1
1Department of Earth and Environmental Sciences, 31 Williams Drive, Lehigh University,
Bethlehem, Pennsylvania 18015-3188 USA
2Smithsonian Environmental Research Center, P.O. Box 28, Edgewater, Maryland 21037 USA
Abstract. Solar ultraviolet radiation (UVR) may have beneficial as well as detrimental
effects on living systems. For example, UV-B radiation (280?320 nm) is generally dam-
aging, while UV-A radiation (320?400 nm) may cause damage or stimulate beneficial
photorepair of UV-B damage. The nature of both direct and indirect effects of UVR in
nature depends on both the photon flux density and the spectral composition of the radiation
incident on aquatic organisms across environmental UVR gradients in space (depth, trans-
parency, elevation) and time (diel, seasonal, interannual). Here we use the common and
widespread freshwater cladoceran Daphnia pulicaria as a model organism to demonstrate
the potential importance of these wavelength-specific effects of UVR to the ecology of
aquatic organisms. UVR-exposure experiments are used to manipulate both natural solar
and artificial UVR sources to examine the beneficial as well as detrimental effects of
different wavelengths of UVR. Changes in the spectral composition of solar radiation are
also examined along several natural environmental gradients including diel gradients, depth
gradients, and dissolved organic carbon (DOC) gradients. The implications of variation in
the spectral composition of UVR for aquatic organisms are discussed.
The first biological weighting function (BWF) for a freshwater cladoceran is presented
here. It demonstrates that the shortest UV-B wavelengths in sunlight are potentially the
most damaging per photon. However, due to the greater photon flux density of longer
wavelength UVR in sunlight, the net potential damage to Daphnia in nature is greatest for
the longer wavelength UV-B and shorter wavelength UV-A radiation in the 305?322 nm
range. Overall the contribution of UV-B to the total mortality response of Daphnia exposed
to full-spectrum solar radiation for 7 h on a sunny summer day is 64% while UV-A con-
tributes 36%. The BWF for Daphnia is used with the transmission spectrum for Mylar D
to demonstrate that Mylar D cuts out only about half of the damaging UVR in sunlight.
Following exposure to damaging UV-B, Daphnia exhibits a dramatic increase in survival
in the presence of longer wavelength UV-A and visible radiation due to the stimulation of
photoenzymatic repair. We present data that demonstrate the importance of both atmospheric
ozone and DOC in creating strong environmental gradients in the intensity (irradiance) and
spectral composition of solar UVR in nature. The light-absorbing component of DOC,
chromophoric dissolved organic matter (CDOM), is particularly important in creating depth
refugia from damaging UV-B in freshwater ecosystems. CDOM may also cause intense
variations in the ratio of potentially beneficial UV-A to detrimental UV-B radiation to which
aquatic organisms are exposed. In addition to changes in atmospheric ozone, future changes
in CDOM related to climate change or other environmental disturbances may substantially
alter the underwater exposure of a variety of aquatic organisms to different wavelengths
of solar UVR.
Key words: biological weighting function; Daphnia pulicaria; dissolved organic carbon; envi-
ronmental gradients; lake optics; ozone depletion; spectral composition of solar radiation; ultraviolet
radiation; UV-A and UV-B, relative damage to aquatic organisms; zooplankton, tolerance of UV
radiation.
INTRODUCTION
Solar radiation influences living organisms in many
different ways depending on its spectral composition.
Longer wavelengths are required for vision in animals
and photosynthesis in plants, while shorter wavelength
Manuscript received 13 July 2000; revised 29 January 2001;
accepted 23 February 2001.
3 E-mail: cew@lehigh.edu
UV-B radiation (280?320 nm) is damaging to living
organisms. Ozone selectively absorbs the shorter wave-
length UV-B radiation, leading to essentially no sun-
light below 290 nm reaching the earth?s surface. In
recent decades biologically damaging UV-B has been
increasing at temperate as well as polar latitudes in
response to stratospheric ozone depletion (Kerr and
McElroy 1993, Madronich 1994). Recent modeling ef-
forts indicate that ozone depletion will continue well
1844 CRAIG E. WILLIAMSON ET AL. Ecological ApplicationsVol. 11, No. 6
into the new century, and recovery will be aggravated
by increases in greenhouse gases (Shindell et al. 1998).
In aquatic habitats chromophoric dissolved organic
matter (CDOM, the light-absorbing component of dis-
solved organic carbon) strongly and selectively absorbs
ultraviolet radiation (UVR), providing organisms with
an additional level of protection from UVR damage
(Scully and Lean 1994, Morris et al. 1995). CDOM
quantity and quality are in turn altered by disturbances
such as climate change and acidification of inland wa-
ters, which may further elevate levels of potentially
damaging UVR for organisms in underwater habitats
(Schindler et al. 1996, Williamson et al. 1996, Yan et
al. 1996). Photobleaching by sunlight may also sub-
stantially increase water transparency to UVR (Morris
and Hargreaves 1997).
One of the most striking pieces of evidence that UV
radiation has been an important selective force in eco-
logical communities is the widespread presence of be-
havioral, physiological, and molecular mechanisms to
avoid, reduce, or repair UVR damage (Williamson and
Zagarese 1994, Zagarese and Williamson 1994, Roy
2000). In spite of the prevalence of numerous UVR
defenses, the detrimental effects of UVR have been
documented in a wide variety of organisms in both
terrestrial and aquatic ecosystems ranging from bac-
teria, protists, and plants to both invertebrate and ver-
tebrate animals (Young et al. 1993, Williamson et al.
1994, Ha?der and Worrest 1997). While shorter wave-
length UV-B is generally damaging, longer wavelength
UV-A radiation (320?400 nm) may be either detri-
mental or beneficial. For example, UV-A radiation may
cause substantial inhibition of growth or survival in
bacteria (Sommaruga et al. 1997), phytoplankton
(Smith et al. 1992, Pre?zelin et al. 1994), protozoa (Ochs
and Eddy 1998), zooplankton (Williamson et al. 1994),
and fish (Williamson et al. 1997). On the other hand,
UV-A may have beneficial effects by stimulating pho-
toenzymatic repair (Sutherland 1981, Mitchell and Kar-
entz 1993, Quesada et al. 1995). Photoenzymatic repair
(PER, also referred to as ??photoreactivation,?? or ??pho-
torepair??) is a light-mediated molecular process that
repairs UV-B-damaged DNA. Fish larvae benefit from
PER as well as from UV-A receptors in their retinas
that enhance their foraging efficiency on zooplankton
prey (Loew et al. 1993, Browman et al. 1994). Blue
light may stimulate the production of photoprotective
compounds in plants (Adamse et al. 1994). UV-A in-
hibits photosynthesis (Neale 2000) but is also photo-
synthetically active since most light-harvesting pig-
ments also absorb in the UV-A region (Vernet et al.
1989). The lack of good quantitative information on
the contrasting detrimental and beneficial effects of the
different wavelengths of solar radiation in nature makes
it difficult to predict the net response of organisms to
changing environmental UVR.
In order to understand how natural communities re-
spond to environmental variations in UVR, information
on how organisms respond to the different wavelengths
is essential (Smith et al. 1980, Caldwell et al. 1986,
Cullen and Neale 1997). Two primary approaches have
been used in the past to examine the response of living
organisms to different wavelengths of UVR. Mono-
chromatic approaches expose organisms to a narrow
wavelength band of UVR and use the response to de-
velop an action spectrum (Setlow 1974, Storz and Paul
1998). Polychromatic approaches on the other hand
expose organisms to full-spectrum radiation and pro-
gressively remove the shorter wavelengths in multiple
treatments to develop biological weighting functions
(Neale 2000). The polychromatic approach is generally
considered to be more ecologically relevant (Caldwell
et al. 1986), and is the approach that we use here.
Both monochromatic and polychromatic approaches
confirm that per photon, biological damage increases
with decreasing wavelength in the UVR range. How-
ever, at the same time, fewer photons of this shorter
wavelength solar UVR reach Earth?s surface, due in
large part to absorption by atmospheric ozone. This
tradeoff between higher photon-specific damage and
decreasing photon flux density with decreasing wave-
length leaves open the central question of which wave-
lengths of sunlight actually contribute the most to UVR
damage in nature. This is an especially important con-
sideration in aquatic systems where CDOM creates in-
tense gradients in the spectral composition of sunlight
and in the ratio of beneficial to detrimental wavelengths
of UVR.
Here we use Daphnia pulicaria as a model organism
to look specifically at what wavelengths of solar ra-
diation are most damaging. We also examine the po-
tentially beneficial effects of longer wavelength UV-A
in stimulating photorepair of UV damage. Daphnia was
chosen as the experimental organism because it forms
a particularly important link in many pelagic freshwater
food webs. It is an important food item in the diet of
planktivorous fish, and it exerts strong grazing pres-
sures on phytoplankton and bacteria. In addition, Daph-
nia, like many aquatic organisms ranging from phy-
toplankton to fish, has the ability to photorepair UVR
damage. There are clear ecological analogues to Daph-
nia in other ecosystems that are exposed to UVR such
as the phytophagous insects in terrestrial and lotic eco-
systems, and copepods in marine ecosystems.
Data are also presented here on the variability of
both the intensity (irradiance) of UV-A and UV-B as
well as the ratio of UV-A to UV-B radiation along
several common environmental gradients (CDOM gra-
dients, depth gradients, diel gradients). Although there
is variation in the wavelength-specific effects of UVR
among organisms and ecosystems, the general princi-
ples demonstrated here with Daphnia apply to other
organisms?UVR can have both beneficial and detri-
mental effects in nature, and the net effect will depend
on the spectral composition as well as the response of
the organisms exposed.
December 2001 1845UVR SPECTRAL COMPOSITION AND IMPACT
FIG. 1. Diagram of the solar phototron used
to expose Daphnia to different wavelengths of
natural sunlight.
Most organisms respond to damaging UVR through
some combination of behavioral avoidance, photopro-
tection, and molecular repair (Worrest 1982, Zagarese
and Williamson 1994). Here we look only at the two
latter responses, photoprotection and molecular repair,
the sum of which we define as ??UV tolerance.?? More
specifically we use two novel instruments to examine
the effects of the quality (spectral composition) as well
as the quantity (irradiance) of UVR on Daphnia puli-
caria in a series of experiments under both natural and
artificial solar-radiation regimes using polychromatic
approaches. First we use a solar phototron with long-
wave pass cutoff filters and natural sunlight to examine
wavelength-specific responses to solar radiation and
develop a biological weighting function (BWF) for
Daphnia. From this BWF we determine which wave-
lengths of sunlight cause the most net damage during
a midday exposure to UVR, as well as the total bio-
logically weighted exposure (H*), which is essentially
the UVR dose. Second, we use a UV-lamp phototron
that permits us to both manipulate UVR irradiance and
to separate the effects of potentially beneficial longer
wavelength radiation from the damaging shorter wave-
length UV radiation. We use this latter approach to
assess the importance of both the level of damaging
radiation and the effectiveness of photoenzymatic re-
pair (PER) in determining net damage from UVR. Un-
derstanding the importance of PER is critical because
in the presence of PER the principle of reciprocity
(where response is dose dependent rather than dose-
rate dependent [Luckiesh 1930]) is invalid and one
must then be very cautious about applying weighting
functions to diverse radiation regimes (Cullen and Nea-
le 1997). Lastly we examine and discuss natural en-
vironmental variations in UV-A, UV-B, and UV-A:UV-
B ratios in aquatic ecosystems and the implications of
these gradients in UVR for the ecology of organisms
such as Daphnia.
METHODS
Solar-phototron experiment: What wavelengths
are damaging?
Daphnia pulicaria were collected from Dutch
Springs in Bethlehem, Pennsylvania, USA, a UV-trans-
parent lake (1% of surface UV 320-nm radiation pen-
etrates to ;8 m in summer) with a maximum depth of
30 m. The Daphnia were exposed to various wave-
lengths of natural sunlight in the solar phototron (Fig.
1) in order to assess the wavelength-specific nature of
their response to net UV radiation (UVR) damage. The
Daphnia were collected with a 202-mm mesh net and
brought back to the laboratory where they were fed and
acclimated to the experimental temperature of 208C
overnight.
The following morning 10 adult Daphnia (egg bear-
ers, or individuals the size of egg bearers) were placed
in each of five replicate dishes for each of five light-
filter treatments. The dishes were shallow (1.7 cm deep,
4.8 cm inside diameter, ;30 mL capacity) to reduce
potential shading effects, and painted black on the out-
side to remove all solar radiation except that which
came down through the filter that was placed on top
of the dish as a cover. Quartz disks and Schott long-
wave pass cutoff filters (50-mm diameter) (Schott
Group, Mainz, Germany) with four different wave-
length cutoffs were used to manipulate the wavelengths
of sunlight to which the organisms in each treatment
were exposed. The Schott filters had sharp cutoffs with
50% transmittance at 304, 326, 370, and 404 nm. The
quartz treatment permitted exposure to essentially full
solar radiation. We found that solar irradiance de-
creased much more rapidly below 293 nm (an order of
magnitude per nanometer) and thus used 293 as the
nominal ??cutoff?? for the quartz treatments. The re-
sulting cumulative energy spectra to which the organ-
isms were exposed are given in Fig. 2.
The dishes with filter covers were suspended on a
1846 CRAIG E. WILLIAMSON ET AL. Ecological ApplicationsVol. 11, No. 6
FIG. 2. Solar energy spectra for quartz and
Schott long pass cutoff filter treatments. From
left to right, the spectra are for quartz, 305-,
326-, 370-, and 404-nm filter treatments. Data
are from the 7-h experimental period for the
solar phototron experiment on 10 July 1998. See
Methods for details.
rack in a large flat tray with water circulating under-
neath and on all sides from a temperature-controlled
water bath (Fig. 1) to maintain the experimental tem-
perature at 208C. The incubation period lasted 7 h be-
ginning at 0930 and ending at 1630 Eastern daylight
time EDT on 10 July 1998. The day was very sunny
with only occasional short periods of passing clouds.
At the end of the experiment the dishes were transferred
to an incubator with a 12:12 light:dark cycle (two 40-
W cool-white fluorescent bulbs). Mortality in each dish
was scored 42 h after the end of the exposure period.
The UVR exposure in each light-filter treatment was
estimated from measured solar-irradiance data, the
transmission spectra of the filters, and a solar-radiation
model. Solar irradiance was measured continuously
throughout the 7 h exposure period at 1?2 s intervals
with a carefully leveled profiling ultraviolet radiometer
(PUV 501B, Biospherical Instruments, San Diego, Cal-
ifornia, USA; 1997 solar calibration coefficients in air)
placed within a few meters of the solar phototron. The
PUV is a medium-bandwidth radiometer that measures
photosynthetically active radiation (PAR, 400?700 nm)
as well as UVR in four bandwidths (305, 320, 340, and
380 nm). The UVR bands have a bandwidth of 8?10
nm measured as the full width at half maximum re-
sponse (FWHM, the range between the two wave-
lengths at which response is 50% of the peak response).
The UV irradiance data (in units of mW?cm22?nm21)
were integrated over the 7-h exposure period to esti-
mate total energy exposure (in units of J?m22?nm21) at
each PUV wavelength. These integrated data were in
turn used to generate a solar energy exposure spectrum
(280?500 nm, at 1-nm intervals) by fitting a solar spec-
trum to them. The shape of the solar spectrum was
generated using a solar radiation model (Biospherical
Instruments: RT Basic radiative transfer model) that
quantifies the propagation of different wavelengths of
solar radiation through the atmosphere to permit esti-
mation of the shape of the solar spectrum on a given
date and location. A previously developed protocol for
fitting full-spectrum data to medium-bandwidth instru-
ment data was used (Kirk et al. 1994, Morris and Har-
greaves 1997). The transmission spectra of the filters
in each treatment were multiplied by the solar energy
exposure spectrum to produce the wavelength-specific
energy spectra for each filter treatment with 1-nm res-
olution (Fig. 2).
Biological weighting function (BWF) approach
and model
The basic concept behind a biological weighting
function is that the impact of solar radiation depends
on its spectral composition: different wavelengths have
different effects. A BWF is used to estimate the effects
of different wavelengths on a given organism or pro-
cess by generating a function that provides a weighting
coefficient for each individual wavelength. Either a
general shape is assumed for the function, or multi-
variate statistical approaches are used to generate more
complex functions. Here the data from the solar-pho-
totron experiment were used to develop a BWF for net
damage (combination of damage and repair) in Daph-
nia pulicaria. We fit a simple exponential function to
the Daphnia mortality data following a modification of
previously described methods (Rundel 1983, Cullen
and Neale 1997, Kouwenberg et al. 1999b). This ap-
proach is based on the concept that there are two factors
that determine the impact of UVR on a living organ-
ism?the amount of radiation at each wavelength, and
the effectiveness of the energy of the photons at each
wavelength. Thus the model expresses net biologically
weighted or effective exposure (H*, unitless) as a func-
tion of total cumulative energy exposure at each wave-
length (H(l), in joules per square meter per nanometer)
and the biological weighting coefficient for each wave-
length (?H(l), in [joules per square meter]21), integrated
over the wavelength range of interest. We chose the
280?500 nm wavelength range as being comprehensive
December 2001 1847UVR SPECTRAL COMPOSITION AND IMPACT
FIG. 3. Diagram of UV-lamp phototron used to expose
Daphnia to different intensities of damaging UV-B radiation
and to manipulate the longer-wavelength photorepair radia-
tion (PRR). Positions on the wheel where dishes are exposed
or unexposed to PRR are indicated.
to any potentially damaging effects of solar radiation
to zooplankton:
500
H* 5 ? (l) 3 H(l) 3 Dl. (1)
O H
l5280
When the values of ?H(l) are plotted against wave-
length they comprise what is known as a ??biological
weighting function.?? They are estimated from the fol-
lowing model:
{2[m 1m ?(l2300)]}1 2
? (l) 5 C 3 e 1 mH c (2)
where m1 is the natural log of the weight at 300 nm,
m2 is the slope of the BWF, C is a proportionality con-
stant ([J/m2]21), here equal to 1, and mc is a constant
that accounts for wavelength-independent effects. Bi-
ologically effective exposure at a given wavelength
(H*
l
; nm21) is equal to the product of the energy ex-
posure at each wavelength and the biological weighting
coefficient for that wavelength:
H* 5 ? (l) 3 H(l).
l H (3)
When is integrated over the full wavelength rangeH*
l
(Eq. 1) the result is H*.
The model that we used for estimating mortality due
to H* was based on the standard exponential equation
for population growth where N0 was the starting num-
ber of individuals and Nt was the number of survivors
at the end of the experimental time period:
2HN 5 N e *.t 0 (4)
When modified to express the results as percentage
mortality as was done here, this becomes
21 2H100 3 (1 2 N 3 N ) 5 100 3 (1 2 e *).t 0 (5)
These relationships assume that the natural log of sur-
vivorship is inversely proportional to total biologically
weighted UVR exposure (H*), and that reciprocity
holds. Total biologically weighted or effective expo-
sure is roughly the same thing as dose except that the
term ??dose?? is usually reserved for that exposure
which is actually absorbed. The ??reciprocity principle??
states that the response to a given UVR dose is inde-
pendent of dose rate (Luckiesh 1930, Cullen and Neale
1997). So for example, if reciprocity holds, an organ-
ism will respond in a similar way to a given UVR dose
whether that dose is received in a single sunny day, or
over a period of several cloudy days.
The BWF was estimated numerically given Eqs. 1?
4 using a Marquardt nonlinear least-squares algorithm
(SAS Institute 1996). The fitting procedure determined
the best fit values for m1, m2, and mc given spectral
exposure and observed mortality (as discussed in Solar-
phototron experiment. . . , above). Each replicate dish
was considered an independent observation (n 5 25).
The 95% confidence interval for the ?H estimates de-
rived from Eq. 2 were determined from the variance
and covariance of the fitted m1, m2, and mc using prop-
agation of errors (Bevington 1969).
UV lamp phototron experiments: What is the
importance of UV damage and photorepair?
The UV-lamp phototron (Fig. 3) is a novel instrument
designed to expose organisms to damaging UV-B ra-
diation in the presence or absence of beneficial longer
wavelength photoreactivating radiation (PRR). The
standard exposure period is 12 h, after which the sur-
vival (or growth rate for rapidly reproducing organ-
isms) is monitored until the dark controls drop below
90% survival or for a total of 5 d at 208C (5endpoint).
Percentage survival (or growth rate) at the endpoint is
used to measure UV tolerance and its two components:
PER (photoenzymatic repair) and DRPP (dark repair
and photoprotection). We define ??UV tolerance?? as the
percentage survival at the endpoint in the treatment
with both UV-B and PRR (1PRR treatment). In the
absence of PRR (2PRR treatment), there is little or no
PER so that organism survival following UV-B ex-
posure will be a function of dark repair processes and
photoprotective ability, giving us an estimate of what
we define here as DRPP. It is important to recognize
that substantial photoprotection may be present, and
dark repair may be going on even if there is no survival
in the absence of PRR (i.e., zero DRPP as defined here).
We are using percentage survival at a given UV ex-
posure level as the operational definition of DRPP to
measure the consequences of the dark repair and pho-
1848 CRAIG E. WILLIAMSON ET AL. Ecological ApplicationsVol. 11, No. 6
FIG. 4. Energy spectra for the photorepair radiation (PRR;
thin line) and UV-B (medium line) sources integrated over a
12-h exposure period in the UV-lamp phototron compared to
the energy spectrum for the 7-h sunlight exposure under
quartz in the solar phototron (thick line).
toprotection rather than directly measuring the physi-
ological and molecular repair processes. PER in turn
is measured as the difference in survival in the treat-
ments with (1PRR) and without (2PRR) photorepair
radiation. The design of the phototron also permits the
investigator to manipulate both the wavelength and the
intensity (irradiance) of exposure to both UV-B and
PRR in any one experiment.
The UV-lamp phototron consists of a UV-B lamp
(Spectronics XX15B) suspended 24 cm above the sur-
face of a rotating black acrylic wheel that has 40 open
holes in it in which 1.8 cm deep, 4.6 cm inside diameter,
;30 mL capacity quartz petri dishes are placed (Fig.
3). The UV-B lamp is covered with a fresh sheet of
cellulose acetate at the start of each 12-h exposure to
cut out the wavelengths of UV-B shorter than ;295
nm and thus better simulate natural solar radiation.
Each dish is surrounded by a 2.5-cm-tall black collar
of polyvinyl chloride to minimize exposure to stray
radiation. Quartz lids cover the dishes. Stainless-steel
mesh screens are placed on top of the dishes to ma-
nipulate the intensity of UV-B exposure. The wheel
covers an enclosed box that contains the lamps that
emit PRR consisting of visible, UV-A, and a small
amount of UV-B radiation. The PRR bulbs are two 40-
W cool-white fluorescent bulbs and two 40-W Q-Panel
340 bulbs located 32 cm below the bottom of the dishes.
A black skirt around the acrylic wheel prevents stray
PRR from escaping the box. By adding black disks to
cover the holes under the dishes the presence or absence
of the PRR from below can be manipulated. The box
is ventilated with a high-rpm thermostatically regulated
fan to prevent heat build up. The wheel rotates the
specimens horizontally at 2 rpm to provide uniform
exposure of all dishes and simulate the variable inten-
sities of UVR that occur with mixing in the surface
waters of lakes. The entire apparatus is placed inside
a growth chamber at a constant temperature of 208C.
The spectral composition of the damaging radiation
from above and repair radiation from below the wheel
were measured at 1-nm resolution with a custom-made
spectral radiometer. Measurements were made at 14
different positions of the rotating wheel for the UV-B
lamp radiation and 8 different positions of the rotating
wheel for the repair radiation. Values were integrated
over these different positions to account for the rotation
of the wheel in order to estimate spectral exposures.
In addition, a laboratory spectrophotometer was used
to measure and account for the degradation of the cel-
lulose acetate over a 12-h period of exposure to the
UV-B lamp. The resulting energy spectra from the UV-
B and PRR lamps were compared with the solar energy
spectrum from the solar-phototron experiment (Fig. 4).
The spectral radiometer that was used to make these
measurements was custom made by Patrick J. Neale.
It consists of a scanning monochromator (model SP
300i, Acton Research Corporation, Acton, Massachu-
setts, USA) with a UV-sensitive PMT (1P28 photo-
multiplier tube) connected to a 3-m fiber-optic cable
and a cosine-corrected flat diffuser collector. The re-
sponse was calibrated for wavelength using a mercury
lamp and for irradiance using a 1000-W NIST-traceable
standard lamp (National Institute of Standards and
Technology [U.S. Department of Commerce], Gaithers-
burg, Maryland, USA).
For the current experiments five 2?3 d old (imma-
ture) Daphnia pulicaria from a laboratory culture
(originating from Dutch Springs) were placed in each
of five replicate dishes in eight different treatments?
four levels of UV-B exposure in the presence and ab-
sence of PRR. The experiment was run in December
1999. Preceding the experiment the Daphnia culture
was kept at 208C and a 12:12 L:D cycle from two 40-
W cool-white fluorescent bulbs and fed Cryptomonas
reflexa. The treatment levels were 11, 26, 34, and 55
kJ/m2 of UVR (280?400 nm) from the UV-B lamp.
Each treatment level had five dishes exposed to PRR
and five not exposed to PRR. The exposure period in
the phototron was 12 h. Survival was recorded under
a dissecting microscope in each dish every day for five
days following the start of the exposure period. The
endpoint data from day 5 were corrected for dark con-
trol survival (96%) with a modification of Abbott?s
equation (Newman 1995, Williamson et al. 1999), arc-
sine square root transformed, and a two-way ANOVA
performed on the transformed data where UV exposure
level and PRR level were the two factors.
Environmental variation in UV radiation
Selective absorption of short wavelength UVR by
ozone in the atmosphere and by dissolved organic car-
bon (DOC) in aquatic ecosystems creates strong gra-
dients in UVR in nature. Some of the more obvious
patterns are the diel and annual variations in UVR re-
lated to changes in sun angle, and the decrease in UVR
with increase in depth in the water column of aquatic
December 2001 1849UVR SPECTRAL COMPOSITION AND IMPACT
FIG. 5. Percentage mortality of Daphnia pulicaria 42 h
after the end of the 7-h exposure in the solar phototron, with
filter cutoff allowing 50% transmittance. Observed mortality
is given as means with 95% confidence intervals. Modeled
mortality is based on the biological weighting function fit by
iteration to the observed mortality data. Full-spectrum
(quartz) treatment is plotted at 293 nm, as below this wave-
length solar irradiance declines by an order of magnitude per
nanometer.
FIG. 6. Biological weighting function (BWF) for Daphnia
pulicaria (thick line). Thin lines represent 61 SE error above
and below the BWF. The BWF equation (Eq. 2) here is ?H(l)
5 7.5 3 1025(e(20.132(l2300) 1 6.6 3 1028).
ecosystems. Perhaps more interesting and less widely
recognized is the fact that this selective absorption of
shorter wavelength UVR will lead to wide variations
in the ratio of longer wavelength (often beneficial) UV-
A radiation compared to the shorter wavelength (usu-
ally damaging) UV-B radiation. Here we present data
on UV-A (380 nm), UV-B (320 nm), and UV-A:UV-B
ratios in lake ecosystems.
Our UV group in the Department of Earth and En-
vironmental Sciences at Lehigh University (Bethlehem,
Pennsylvania, USA) has been collecting vertical pro-
files of underwater UVR since1993 with Biospherical
PUV-501 and PUV-501B submersible profiling ultra-
violet radiometers. We have also had a ground-based
ultraviolet (GUV) radiometer installed in a weather sta-
tion at our field research site at Lake Lacawac in north-
eastern Pennsylvania (USA) since 1993. These instru-
ments record UV irradiance for 305, 320, 340, and 380
nm, as well as photosynthetically active radiation
(PAR; 400?700 nm). Diffuse attenuation coefficients
(Kd) are obtained for the water column of lakes from
the slope of the linear regression of the natural loga-
rithm of down-welling irradiance (Ed) vs. depth over
depth intervals where the relationship is linear. Past
intercomparisons and optical modeling have shown that
although the PUV and GUV have a moderate band-
width (8?10 nm FWHM), Kd measured with these in-
struments compare favorably with those measured with
spectral radiometers of 1?2 nm bandwidths. Thus they
can be used to estimate irradiance over the entire UVR
spectrum within a few percent (Kirk et al. 1994, Lau-
rion et al. 1997). Here we use data from our UV ar-
chives to illustrate the nature of environmental gradi-
ents in UV-A (380 nm) and UV-B (320 nm) as a func-
tion of depth in the water column of optically different
lakes. We also present data on variation in the UV-A:
UV-B ratio as a function of depth in the water column,
as a function of time of day (sun angle), and as a func-
tion of DOC concentration among lakes sampled in a
broad survey of lakes in both the northern and southern
hemispheres (Morris et al. 1995).
RESULTS
Solar-phototron experiment
Mortality was highest (averaging 71%) in Daphnia
exposed to full-spectrum solar radiation for 7 h in the
quartz treatment of the solar phototron experiment (Fig.
5). When the shorter wavelength UV-B (,305 nm) was
removed, the average mortality decreased to 57%, and
when most of the UV-B (,326 nm) was removed mor-
tality further decreased to an average of 36%. When
the medium and short wavelength UV-A (,370 nm)
was also removed the mortality declined to 23% (Fig.
5). When essentially all of the ultraviolet radiation
(UVR) was removed in the 404-nm treatment the mor-
tality averaged only 4%.
Biological weighting function
The biological weighting function (BWF) for Daph-
nia demonstrates a strongly increasing energy-specific
effect with decreasing wavelength (Fig. 6). The shortest
wavelength UV-B radiation in sunlight is on the order
of a thousand times more damaging to Daphnia than
is the longest wavelength UV-A per unit energy. The
photon-specific effect is similar to the energy-specific
effect (but not shown here). However, incident solar
energy also shows a strong decrease with decreasing
wavelength within the UV-B range (Fig. 7). In order
to estimate which wavelengths of sunlight are most
damaging to Daphnia, the BWF (which gives us en-
ergy-specific damage rates) was multiplied by the solar
energy spectrum (which tells us how much energy sun-
light has at each wavelength). The result is biologically
1850 CRAIG E. WILLIAMSON ET AL. Ecological ApplicationsVol. 11, No. 6
FIG. 7. Biologically effective exposure
( ; thin line) as estimated from the 7-h solarH*
l
phototron exposure under quartz (thick line) and
the biological weighting function, BWF (me-
dium line). H(l) is wavelength; ?H(l) is the bi-
ological weighting coefficient for each wave-
length.
FIG. 8. Biologically effective exposure ( ) estimates forH*
l
the solar phototron quartz treatment (thick line), the UV-B
lamp (medium line), and the photorepair radiation (PRR)
lamps (thin line) in the UV-lamp phototron. These valuesH*
l
were estimated by multiplying the biological weighting func-
tion for Daphnia in Fig. 6 by the lamp and solar spectra in
Fig. 4.
effective exposure at each wavelength ( , Fig. 7).H*
l
This shows us that the most biologically damaging ra-
diation (for Daphnia under these conditions) is in the
305?322 nm range (bandwidth measured as the full
width at half maximum response, FWHM; i.e., the
range between two wavelengths at which response is
50% of the peak response). The BWF also indicates
the relative contribution of UV-B and UV-A exposure
to mortality for the solar phototron. Overall UV-B con-
tributed 64% of the total response to UV in the full-
spectrum quartz treatment while UV-A contributed
36%.
Nonlinear fitting of the parameters of the BWF for
Daphnia resulted in values of 9.49 6 0.97 (mean 6 1
SE) for m1, 0.132 6 0.053 for m2, and 6.6 6 2.6 3 1028
for mc (see Eq. 2). The r2 for the fit of the model was
0.72, but 23% of the variance was replicate variance.
In terms of the between-treatment variance (the only
variance that can be explained by the model) the model
accounted for 94% of the variance. The statistical var-
iation among replicates (Fig. 5) does leave the exact
shape of the BWF uncertain. The standard errors es-
timated through propagation of errors are particularly
high below 305 nm and between 335?355 nm, wave-
lengths that contribute little to effective solar exposure
(Fig. 6). However, the BWF is fairly well constrained
over the 305 to 335 nm band (centered on 320 nm),
which contributes the most to solar exposure. The fit
required addition of a ??constant?? weight (mc), which
dominates the BWF at wavelengths . 370 nm. This
constant provides a simple approximation of the net
effect of exposure to long wavelength UV-A and visible
light.
The weighting function probably does have some
wavelength dependence at longer UV-A and visible
wavelengths but the slope is much shallower than in
the UV-B and short wavelength UV-A. If we assume
that the variance is similar among treatments and cal-
culate a pooled variance over all treatments (sample
variance as estimated by one-way ANOVA of per-
centage survival as a function of spectral treatment,
with 20 degrees of freedom), then the standard error
of the mean in each treatment is 63% survival. The
response to the long wavelength UV-A and visible light
(370-nm cutoff) treatment is then significant but the
response to the visible-only treatment remains statis-
tically indistinguishable from zero.
We also used the Daphnia BWF to compare the bi-
ologically effective UV exposures (H* values) in the
solar phototron to those in our UV-lamp phototron to
assess how the biologically effective exposures in the
UV-lamp phototron compared to more natural solar ra-
diation. To do this we multiplied the BWF (Fig. 6) by
each energy spectrum in the UV-lamp phototron (Fig.
4). For the 7-h solar exposure on 10 July the value of
H* was 1.00 (area under the solar curve in Fig. 8, 280?
December 2001 1851UVR SPECTRAL COMPOSITION AND IMPACT
FIG. 9. Contribution of photoenzymatic repair (PER) vs.
dark repair and photoprotection (DRPP) to overall UVR tol-
erance of Daphnia pulicaria. DRPP and tolerance are both
expressed as percentage survival (means and 1 SE), while
PER is the difference between these two. The data are end-
point data on day 5 following a single 12-h exposure to dam-
aging ultraviolet radiation (UVR) in the UV-lamp phototron
in the presence and absence of photorepair radiation (PRR).
400 nm). To extend these data to a full-day equivalent,
we assumed that H* follows the same trend as 320 nm
from dawn to dusk (Fig. 12B). This was necessary
because different wavelengths follow different trends
over the diel cycle. The result was that this 7-h period
represents 77% of the solar radiation in a full day of
UV exposure at this time of year, giving a value of H*
5 1.30 for a full day of solar UVR under quartz.
We estimated that during a 12-h exposure to the full-
strength UV-lamp phototron, Daphnia receives UVR
damage equivalent to 86% of that received on a sunny
day near summer solstice. In the UV-lamp phototron
the biologically effective exposure for a 12-h exposure
to one UV-B lamp gives a value of H* 5 0.94 (Fig. 8,
UV-B lamp curve, 280?400 nm). The repair radiation
from within the phototron box also produces a small
amount of damaging radiation (H* 5 0.18, Fig. 8 PRR
curve, 280?400 nm) that, when added to the UV-B,
gives an H* value of 1.12, which is 86% of 1.3.
UV-lamp phototron experiments
The experiments with the UV-lamp phototron
showed a clear effect of UV exposure level (P , 0.001)
as well as an effect of the addition of longer wavelength
photorepair radiation (P , 0.001) on the survival of
juvenile Daphnia exposed to damaging UVR. The UVR
tolerance of Daphnia (expressed as percentage surviv-
al) decreased from 89% at 11 kJ/m2, to zero in the fully
exposed treatments (55 kJ/m2, Fig. 9). Photoenzymatic
repair (PER) was by far the most important component
of the observed UVR tolerance at all exposure levels.
In the absence of photorepair radiation (2PRR treat-
ments), the DRPP (dark repair and photoprotection)
alone resulted in little or no survival: only 2 out of 100
individuals survived in all four of the 2PRR treat-
ments.
Environmental variation in UV radiation
Here we present data on environmental gradients in
both UVR irradiance and ratio of UV-A (380 nm) to
UV-B (320 nm) radiation. These wavelengths were cho-
sen based on the availability of our extensive PUV
(profiling ultraviolet) and GUV (ground-based UV) ra-
diometer data and the knowledge that solar radiation
in the 320-nm range is the most biologically effective
in causing damage to Daphnia (Fig. 7). For ease of
understanding these wavelengths are referred to here
as ??UV-A?? and ??UV-B,?? respectively.
Underwater UVR varies greatly in aquatic ecosys-
tems as a function of the quality and quantity of dis-
solved organic carbon, DOC. For example in lakes with
low DOC the 320-nm UV-B may penetrate to depths
of 8 m or more, while in lakes with even moderate
DOC concentrations most UV-B is absorbed within ,1
m of the surface (Fig. 10). Thus a major portion of the
mixed layer in low-DOC lakes may be exposed to levels
of damaging UV-B that are 10% or more of surface
levels while in lakes with higher DOC a substantial
refuge from damaging UVR exists within the lower
portion of the mixed layer (Fig. 10).
Within a lake UV-A radiation penetrates substan-
tially deeper than does UV-B (Fig. 10). At the same
time, the ratio of beneficial UV-A to damaging UV-B
radiation remains very low (,10) throughout the high
UVR mixed layer of low-DOC lakes, but increases to
much higher levels within the top meter of moderate-
DOC lakes (Fig. 11). This means that lakes with higher
levels of DOC not only have less of the damaging UV-
B radiation, but also proportionally more beneficial
UV-A.
The variation in UV-A:UV-B ratios is even more
evident across lakes with a range of DOC concentra-
tions. For example, at a depth of 1 m across a range
of lakes that we sampled as part of a previous study
(Morris et al. 1995), UV-A:UV-B ratios vary over a
thousandfold (Fig. 12A). UV-A:UV-B ratios can also
vary with changes in sun angle due to the increased
cross section of ozone when the sun is lower in the
sky. The variation in the UV-A:UV-B ratio with sun
angle is not as extreme as it is with changes in DOC
underwater, but it can vary from on the order of 2.5
near midday to 10 or more at dawn and dusk (Fig.
12B). This same sort of variation would be expected
with changes in sun angle and ozone cross section dur-
ing different times of the year or at different latitudes.
DISCUSSION
The above experiments provide the first biological
weighting function (BWF) and estimate of biologically
effective exposure for a freshwater cladoceran. They
demonstrate the potential importance of the spectral
composition of sunlight to the ecology of Daphnia pul-
1852 CRAIG E. WILLIAMSON ET AL. Ecological ApplicationsVol. 11, No. 6
FIG. 10. UV-A (380 nm) and UV-B (320 nm) radiation vs. depth in two northeast Pennsylvania (USA) lakes with differing
dissolved organic carbon (DOC) concentrations. Lake Giles has a DOC concentration of ;1.2 mg/L, while Lake Lacawac
has a DOC concentration of ;4.7 mg/L. Mixing depths represent the shallowest depth at which the thermal gradient is ,18C/
m depth. Data are from June 1999.
FIG. 11. Ratio of UV-A (380 nm) to UV-B
(320 nm) irradiance vs. depth in lakes Lacawac
and Giles (Pennsylvania, USA) from the surface
to the depth of 1% of the 320-nm surface UV.
Mixing depths and dates are as in Fig. 10.
icaria, a widespread and important grazer in pelagic
freshwater food webs. These experiments also dem-
onstrate the tradeoff between the greater energy-spe-
cific damage of short wavelength ultraviolet radiation
(UVR) and the greater photon flux density of the longer
wavelengths of UVR. For this population of D. puli-
caria the implication is that the most important wave-
lengths of sunlight in terms of net damage are those in
the 305?322 nm range. The BWF for Daphnia falls
between those previously obtained for UV-induced
mortality in eggs of another planktonic crustacean, the
marine copepod Calanus finmarchicus, and cod (Gadus
morhua) (Kouwenberg et al. 1999a, b), but was more
similar in slope to that of the copepod (Fig. 13). A
major difference was that Daphnia did exhibit some
sensitivity to UV-A.
The experiments with the UV-lamp phototron dem-
onstrate the key importance of longer wavelength UV-
A and visible light in stimulating photoenzymatic re-
pair of UVR damage. Interestingly, longer wavelength
UV-A may actually have a net negative effect at very
high irradiance levels (Fig. 6). However, this effect is
orders of magnitude less than that for shorter wave-
length UV-B. The lack of a negative effect of UV-A in
the UV-lamp phototron is attributable to the lower lev-
els of UV-A in the phototron compared to solar radi-
ation (Fig. 4). These contrasting effects of different
wavelengths and irradiance levels of UVR combined
with the intense gradients in both beneficial and dam-
aging UVR in the natural environment suggest the need
for more careful attention to the importance of the full
spectral composition of sunlight in studies of UVR ef-
fects.
Changes in atmospheric ozone are likely to alter the
December 2001 1853UVR SPECTRAL COMPOSITION AND IMPACT
FIG. 12. Variation in ratio of UV-A (380 nm) to UV-B
(320 nm) irradiance along (A) environmental gradients in
dissolved organic carbon (DOC) and (B) time. (A) Ratio of
UV-A to UV-B at 1-m depth vs. DOC concentration in a series
of lakes surveyed in North and South America (Morris et al.
1995). The very high ratios are due largely to the fact that
so little UV-B remains at 1-m depth in lakes with higher DOC
concentrations. (B) Irradiance vs. time of day on 26 June
1999 (Eastern daylight savings time, EDST) for different
wavelengths of ultraviolet radiation. The thick line gives the
ratio of UV-A to UV-B throughout the day. Data were col-
lected at Lake Lacawac in northeastern Pennsylvania, USA,
on a clear day at 15-min intervals with a ground-based UV
radiometer (Biospherical GUV).
FIG. 13. Biological weighting function for Daphnia pul-
icaria plotted with weighting functions obtained for the eggs
of a marine calanoid copepod Calanus finmarchicus and eggs
of Atlantic cod (Gadus morhua) by Kouwenberg et al. (1999a,
b).
potential impact of UV-B on Daphnia that remain in
the surface waters during the day. This can be seen by
the fact that the biologically effective exposure curve
(H*l) is constrained by the abrupt decrease in shorter
wavelength UV-B in the solar spectrum (Fig. 7). Daph-
nia often avoid the surface waters of lakes due to visual
predators or potential damage from UVR (Leech and
Williamson 2001). However, up to 20% or more of the
population of Daphnia catawba may be in the surface
mixed layers during July and August in high-UVR
lakes such as Lake Giles, Pennsylvania (USA) (Leech
and Williamson 2000). High UVR in the surface waters
of lakes may also enhance photobleaching of CDOM
(chromophoric dissolved organic matter), the light-ab-
sorbing DOC (dissolved organic carbon) compounds,
with consequent reductions of UVR absorbance and
greater penetration of UVR in the water column (Morris
and Hargreaves 1997). This creates a positive feedback
that may further increase underwater UVR levels, es-
pecially in low-DOC lakes (Williamson et al. 1996).
It is important to note that the function of the UV-
lamp phototron is to separate out photoenzymatic repair
(PER) and dark repair and photoprotection (DRPP)
components of UVR tolerance. The data from the UV-
lamp phototron should not be used to estimate UVR
damage in nature on an energy basis because the spec-
tral composition of the UVR is very different from that
in sunlight (Fig. 4). For example, the ratio of UV-A:
UV-B (380:320 nm) in the UV-lamp phototron is 0.3
in the fully exposed 1PRR (photoreactivating radia-
tion) treatments, while in natural sunlight (solar pho-
totron quartz treatment) it is 2.2. In nature this ratio
generally increases along environmental gradients in
DOC or ozone (Figs. 11 and 12). The hazards encoun-
tered in extrapolating data from artificial light sources
such as UV-B lamps to nature on an energy basis alone
are evident from a comparison of the survival in the
1PRR and 2PRR treatments. In Fig. 9 we plot survival
as a function of only the UV-B lamp energy as this is
what we manipulated. Survival decreases with increas-
ing energy from the UV-B lamps against a background
of UV-A and visible light (Fig. 9). However, survival
was much higher in the presence of PRR (i.e., higher
energy: add 89 kJ/m2 for PRR 280?400 nm to each
UV-B lamp energy exposure on the x-axis of Fig. 9),
than in the absence of PRR (5UV-B lamp exposure
energy on x-axis in Fig. 9). In addition, PER may be
somewhat underestimated in the UV-lamp phototron.
Other experiments that we have carried out have shown
that while PER is substantial with only two PRR lamps
in the phototron, the PER response does not saturate
until somewhere between four and six PRR lamps
(equal numbers of 40-W UV-A 340 nm and cool-white
lamps; M. E. Behum and C. E. Williamson, unpub-
lished data). The negative impacts of UVR demon-
strated here in both the UV-lamp and solar phototrons
are also likely underestimated due to the short-term,
1854 CRAIG E. WILLIAMSON ET AL. Ecological ApplicationsVol. 11, No. 6
acute nature of the experiments, as well as the focus
on only lethal effects.
Our results also clearly demonstrate the complexity
of estimating UVR damage to organisms such as Daph-
nia that have photoenzymatic repair. In the presence of
photorepair, the reciprocity principle is not valid and
the ability to effectively apply BWFs to dissimilar ra-
diation regimes is compromised (Cullen and Neale
1997). The BWF that we generated is the result of the
combination of both damage and repair processes going
on during a sunny midday exposure near summer sol-
stice at a latitude of about 408N. It tells us the wave-
lengths of solar radiation that cause the greatest net
damage (in terms of acute mortality response) in Daph-
nia under these conditions. The presence of PER in
Daphnia means that reciprocity does not apply. In fact,
we have demonstrated the lack of reciprocity in Daph-
nia pulicaria more explicitly in other experiments
(Grad et al. 2001). The lack of reciprocity means that
this BWF cannot be applied effectively to exposure
conditions that differ greatly from this experiment such
as would occur over longer exposure periods at deeper
depths, during deep-mixing periods in the water col-
umn, or on cloudy days. Similarly, the lack of reci-
procity and differences in the radiation regimes in our
UV-lamp vs. solar experiments also means that our
estimates of H* for the UV-lamp experiments must be
interpreted cautiously. For example, even if we use the
48-h survival data for Daphnia in the UV-lamp pho-
totron experiment (analagous to the 42-h solar photo-
tron survival data), the BWF overestimates survival at
the highest UV-B exposure (33% predicted vs. 4% ob-
served survival), and underestimates it at the lowest
UV-B exposure level (69% predicted vs. 96% observed
survival). At the intermediate exposures the BWF over-
estimates survival by 13?15%. Models that parame-
terize damage and repair processes separately will be
necessary to deal with the inability to apply simple
BWFs to a wide range of optical environments in the
presence of PER. This is one of the major long-term
goals of our studies, and one on which one of us has
already begun to develop some of the basic model pa-
rameters (Neale et al., in press).
Knowledge of BWFs is important for the accurate
interpretation of experiments that use various UVR fil-
ters. For example, many investigators use the sharp
cutoff of Mylar D at 316 nm to experimentally estimate
UV-B effects in nature. This cutoff goes right through
the middle of the biologically effective exposure curve
for Daphnia (Fig. 7), so Mylar D really only removes
a portion of the shorter-wavelength UV-B rather than
all of it. Multiplying the transmission spectrum of My-
lar D by the solar energy spectrum for a full day (10
July 1998 data from the solar phototron experiment),
reveals that Mylar cuts out only about 60% of the UV-
B (solar energy at wavelengths ,320 nm). Multiplying
the transmission spectrum of Mylar D by the biolog-
ically effective UVR (H , Fig. 7), reveals that Mylar*
l
D only cuts out about half (56%) of the biologically
damaging solar UVR. This is an important consider-
ation when interpreting experiments in which UVR-
related mortality has been observed in treatments that
were both shielded and unshielded from UV-B with
Mylar D (Williamson et al. 1994, 1997, Zagarese et al.
1994). Total solar UVR impacts may be grossly un-
derestimated when Mylar D is used to manipulate UVR.
The exact percentages that Mylar D cuts out will vary
with the shape of the solar spectrum at different times,
depths, and geographic locations as well as the shape
of the BWF of a given organism. Investigators should
also be aware that there are several other types of Mylar
that are not made to any particular optical specifica-
tions.
The presence of environmental gradients in the ratio
of beneficial to detrimental UVR raises several inter-
esting ecological questions. One question of particular
interest is the potential importance of a threshold level
above which additional UV-A and visible light no lon-
ger add to the enhancement of PER (M. E. Behum and
C. E. Williamson, unpublished data). Does this thresh-
old vary with the level of UV-B damage? Is more UV-
A required to stimulate the repair of higher levels of
UV-B damage? For example, although the UV-A:UV-
B ratios are much lower in the surface waters of low-
DOC lakes (Fig. 11), the absolute amounts of both
damaging UV-B and beneficial UV-A are much higher
(Fig. 10). If there is a threshold UV-A level above
which PER does not increase, then UV-B damage will
likely increase more rapidly as one approaches the lake
surface than if no threshold exists and PER continues
to increase towards the surface with increasing UV-A.
The BWF experiments for Daphnia suggest that at very
shallow depths such as those in the very surface waters
of lakes, longer wavelength UV-A may reach a level
that actually causes net damage.
Another question is how these environmental UVR
gradients influence the interactions of Daphnia with
other abiotic components of aquatic ecosystems. For
example, there is a distinct demographic advantage to
Daphnia that maximize their time spent in the warm
surface waters of lakes (Orcutt and Porter 1983, Stich
and Lampert 1984). Warmer temperatures speed up
rates of growth and reproduction, and reduce the du-
ration of the juvenile life stages that are most vulner-
able to starvation and invertebrate predation. High
CDOM may lead to shallower thermal stratification and
warmer surface-water temperatures in smaller lakes
(Fee et al. 1996). In these same lakes the higher CDOM
will create a refuge from UVR damage (Figs. 10 and
11). In contrast, the surface waters of low-CDOM lakes
are often somewhat cooler and will thus offer less of
a demographic advantage to Daphnia. In these low-
CDOM lakes much of the surface mixed layer may also
be exposed to high levels of damaging UV-B that are
10% or more of surface levels (Fig. 10).
Environmental UVR gradients can also alter the in-
December 2001 1855UVR SPECTRAL COMPOSITION AND IMPACT
teractions of Daphnia with other trophic levels via in-
direct effects. For example, on the one hand photo-
inhibition of phytoplankton by UVR may reduce the
quality or quantity of Daphnia?s food supply (Moeller
1994, De Lange and Van Donk 1997, Hessen et al.
1997). On the other hand UVR may transform recal-
citrant DOC to more labile DOC that stimulates the
microbial loop and thus enhances Daphnia?s food sup-
ply (H. J. De Lange, D. P. Morris, and C. E. Williamson,
unpublished manuscript).
Environmental UVR gradients may also influence the
interaction of Daphnia with its predators. For example,
larval fish have UV-A receptors that are thought to
enhance their ability to forage on zooplankton prey
(Loew et al. 1993, Browman et al. 1994). These same
photoreceptors may also permit the larval fish to avoid
damaging UVR in the surface waters of lakes. This
may be particularly important to larvae of fish such as
yellow perch, which are positively phototactic during
the first months of their lives (Thorpe 1977), yet perish
when exposed to only a few days of solar UVR in the
surface waters of low-DOC lakes (Williamson et al.
1997). Interestingly, Daphnia also has UV photore-
ceptors (Smith and Macagno 1990), and exhibits a neg-
ative phototactic response when exposed to wave-
lengths of UVR shorter than ;360 nm, but a positive
phototaxis to longer wavelengths of sunlight (Storz and
Paul 1998, Leech and Williamson 2001). The ability
of zooplankton and larval fish to detect and respond to
UVR would enable them to use twilight migrations to
increase their exposure to beneficial UVR and thus pro-
long photorepair in UV-B stressed environments (Fig.
12B). Twilight migrations involve upward migrations
at dawn and dusk, with often substantial presence main-
tained in the surface water during the day (Hutchinson
1967). Parallel migration responses in zooplankton and
fish would also alter the vertical overlap between pred-
ators and their prey during these twilight periods, po-
tentially enhancing predation rates (Williamson 1993).
What are the implications of these results for eco-
system management? Human activities have pro-
nounced effects on both CDOM and atmospheric
ozone, the two major regulators of underwater UVR in
pelagic ecosystems. On a global scale, stratospheric-
ozone depletion has led to an increase in UV-B reaching
the earth?s surface, with the effect generally increasing
with latitude in both northern and southern hemi-
spheres. On a more regional scale, increases in tro-
pospheric ozone from pollution can ameliorate these
UV-B increases (Stolarski et al. 1992, Madronich
1998). The effects of damaging UVR are likely to be
most pronounced in aquatic ecosystems with DOC con-
centrations ,1?2 mg/L where UVR penetration is par-
ticularly high (Williamson et al. 1996). The source of
the DOC will also determine its aromaticity and hence
absorptivity of the CDOM content (McKnight et al.
1994). Local to regional changes in climate, acid pre-
cipitation, or watershed use may alter CDOM levels
and hence UVR in aquatic ecosystems (Schindler et al.
1996, Williamson et al. 1996, Yan et al. 1996). Pro-
tection of wetlands and vegetation coverage within the
watershed may also be important to maintaining
CDOM levels (Engstrom 1987, Engstrom et al. 2000,
Williamson et al. 2001). All of these factors will in
turn influence the availability of refugia from damaging
UV-B, as well as the amounts of potentially beneficial
longer wavelength UVR in aquatic ecosystems. Partic-
ular attention should be given to protecting high-ele-
vation alpine lakes. With their small and fragile wa-
tersheds, short growing seasons, and high exposure to
UVR, they are likely to be some of the most sensitive
ecosystems to future changes in UVR.
ACKNOWLEDGMENTS
We thank David Huff and Scott Mastroiani for help in
constructing the solar and UV-lamp phototrons, Susanne
Metzgar for performing numerous experiments that helped to
develop the protocols for the UV-lamp phototron, Will Schenk
for help with the solar phototron experiment, Don Morris for
sharing his DOC data, Robert Moeller for assistance, and
Stuart Schooley for providing access to Dutch Springs. This
work was supported by NSF grants DEB-9509042, DEB-
9740356, and DEB-9973938.
LITERATURE CITED
Adamse, P., S. J. Britz, and C. R. Caldwell. 1994. Amelio-
ration of UV-B damage under high irradiance. II. Role of
blue light photoreceptors. Photochemistry and Photobiol-
ogy 60:110?115.
Bevington, P. R. 1969. Data reduction and error analysis for
the physical sciences. McGraw Hill, New York, New York,
USA.
Browman, H. I., I. Novales-Flamarique, and C. W. Hawry-
shyn. 1994. Ultraviolet photoreception contributes to prey
search behaviour in two species of zooplanktivorous fishes.
Journal of Experimental Biology 186:187?198.
Caldwell, M. M., L. B. Camp, C. W. Warner, and S. D. Flint.
1986. Action spectra and their key role in assessing bio-
logical consequences of solar UV-B radiation change. Pag-
es 87?111 in R. C. Worrest and M. M. Caldwell, editors.
Stratospheric ozone reduction, solar ultraviolet radiation
and plant life. Springer-Verlag, New York, New York, USA.
Cullen, J. J., and P. J. Neale. 1997. Biological weighting
functions for describing the effects of ultraviolet radiation
on aquatic systems. Pages 97?118 in D. P. Ha?der, editor.
The effects of ozone depletion on aquatic ecosystems. Ac-
ademic Press, San Diego, California, USA.
De Lange, H. J., and E. Van Donk. 1997. Effects of UV-B
irradiated algae on life history traits of Daphnia pulex.
Freshwater Biology 38:711?720.
Engstrom, D. R. 1987. Influence of vegetation and hydrology
on the humus budgets of Labrador Lakes. Canadian Journal
of Fisheries and Aquatic Sciences 44:1306?1314.
Engstrom, D. R., S. C. Fritz, J. E. Almendinger, and S. Jug-
gins. 2000. Chemical and biological trends during lake
evolution in recently deglaciated terrain. Nature 408:161?
166.
Fee, E. J., R. E. Hecky, S. E. M. Kasian, and D. R. Cruik-
shank. 1996. Effects of lake size, water clarity, and climatic
variability on mixing depths in Canadian Shield lakes. Lim-
nology and Oceanography 41:912?920.
Grad, G., C. E. Williamson, and D. M. Karapelou. 2001.
Zooplankton survival and reproduction responses to dam-
aging UV radiation: a test of reciprocity and photoenzy-
matic repair. Limnology and Oceanography 46:584?591.
Ha?der, D.-P., and R. C. Worrest. 1997. Consequences of the
1856 CRAIG E. WILLIAMSON ET AL. Ecological ApplicationsVol. 11, No. 6
effects of increased solar ultraviolet radiation on aquatic
ecosystems. Pages 11?30 in D.-P. Ha?der, editor. The effects
of ozone depletion on aquatic ecosystems. Academic Press,
San Diego, California, USA.
Hessen D. O., H. J. De Lange, and E. Van Donk. 1997. UV-
induced changes in phytoplankton cells and its effects on
grazers. Freshwater Biology 38:513?524.
Hutchinson, G. E. 1967. A treatise on limnology. Volume 2.
Introduction to lake biology and the limnoplankton. John
Wiley & Sons, New York, New York, USA.
Kerr, J. B., and C. T. McElroy. 1993. Evidence for large
upward trends of ultraviolet-B radiation linked to ozone
depletion. Science 262:1032?1034.
Kirk, J. T. O., B. R. Hargreaves, D. P. Morris, R. Coffin, B.
David, D. Fredrickson, D. Karentz, D. Lean, M. Lesser, S.
Madronich, J. H. Morrow, N. Nelson, and N. Scully. 1994.
Measurement of UV-B radiation in two freshwater lakes:
an instrument intercomparison. Archiv fur Hydrobiologie
Beihefte Ergebnisse der Limnologie 43:71?99.
Kouwenberg, J. H. M., H. I. Browman, J. J. Cullen, R. F.
Davis, J.-F. St-Pierre, and J. A. Runge. 1999a. Biological
weighting of ultraviolet (280?400 nm) induced mortality
in marine zooplankton and fish. I. Atlantic cod (Gadus
morhua) eggs. Marine Biology 134:269?284.
Kouwenberg, J. H. M., H. I. Browman, J. A. Runge, J. J.
Cullen, R. F. Davis, and J.-F. St-Pierre. 1999b. Biological
weighting of ultraviolet (280?400 nm) induced mortality
in marine zooplankton and fish. II. Calanus finmarchicus
(Copepoda) eggs. Marine Biology 134:285?293.
Laurion, I., W. F. Vincent, and D. R. S. Lean. 1997. Under-
water ultraviolet radiation: development of spectral models
for northern high latitude lakes. Photochemistry and Pho-
tobiology 65:107?114.
Leech, D. M., and C. E. Williamson. 2000. Is tolerance to
UV radiation in zooplankton related to body size, taxon,
or lake transparency? Ecological Applications 10:1530?
1540.
Leech, D. M., and C. E. Williamson. 2001. In situ exposure
to UV radiation alters the depth distribution of Daphnia.
Limnology and Oceanography 46:416?420.
Loew, E. R., W. N. McFarland, E. L. Mills, and D. Hunter.
1993. A chromatic action spectrum for planktonic preda-
tion by juvenile yellow perch, Perca flavescens. Canadian
Journal of Zoology 71:384?386.
Luckiesh, M. 1930. Artificial sunlight, combining radiation
for health with light for vision. D. Van Nostrand Company,
New York, New York, USA.
Madronich, S. 1994. Increases in biologically damaging UV-
B radiation due to stratospheric ozone reductions: a brief
review. Archiv fu?r Hydrobiologie Beihefte Ergebnisse der
Limnologie 43:17?30.
Madronich, S. 1998. Changes in biologically active ultra-
violet radiation reaching the earth?s surface. Journal of Pho-
tochemistry and Photobiology B: Biology 46:5?19.
McKnight, D. M., E. D. Andrews, S. A. Spaulding, and G.
R. Aiken. 1994. Aquatic fulvic acids in algal-rich antarctic
ponds. Limnology and Oceanography 39:1972?1979.
Mitchell, D. L., and D. Karentz. 1993. The induction and
repair of DNA photodamage in the environment. Pages
345?377 in A. R. Young, L. O. Bjorn, J. Moan, and W.
Nultsch, editors. Environmental UV photobiology. Plenum,
New York, New York, USA.
Moeller, R. E. 1994. Contribution of ultraviolet radiation
(UV-A, UV-B) to photoinhibition of epilimnetic phyto-
plankton in lakes of differing UV transparency. Archiv fu?r
Hydrobiologie Beihefte Ergebnisse der Limnologie 43:
157?170.
Morris, D. P., and B. R. Hargreaves. 1997. The role of pho-
tochemical degradation of dissolved organic carbon in reg-
ulating the UV transparency of three lakes on the Pocono
Plateau. Limnology and Oceanography 42:239?249.
Morris, D. P., H. Zagarese, C. E. Williamson, E. G. Balseiro,
B. R. Hargreaves, B. Modenutti, R. E. Moeller, and C.
Queimalinos. 1995. The attenuation of solar UV radiation
in lakes and the role of dissolved organic carbon. Lim-
nology and Oceanography 40:1381?1391.
Neale, P. J. 2000. Spectral weighting functions for quanti-
fying the effects of ultraviolet radiation in marine ecosys-
tems. Pages 73?100 in S. J. de Mora, S. Demers, and M.
Vernet, editors. The effects of UV radiation on marine eco-
systems. Cambridge University Press, New York, New
York, USA.
Neale, P. J., J. J. Fritz, and R. F. Davis. In press. Effects of
UV on photosynthesis of Antarctic phytoplankton: models
and applications to coastal and pelagic assemblages. Re-
vista Chilena de Historia Natural.
Newman, M. C. 1995. Quantitative methods in aquatic eco-
toxicology. Lewis, Boca Raton, Florida, USA.
Ochs, C. A., and L. P. Eddy. 1998. Effects of UV-A (320 to
399 nanometers) on grazing pressure of a marine hetero-
trophic nanoflagellate on strains of the unicellular cyano-
bacteria Synechococcus spp. Applied and Environmental
Microbiology 64:287?293.
Orcutt, J. D., and K. G. Porter. 1983. Diel vertical migration
by zooplankton: constant and fluctuating temperature ef-
fects on life history parameters of Daphnia. Limnology and
Oceanography 28:720?730.
Pre?zelin, B. B., N. B. Boucher, and R. C. Smith. 1994. Marine
primary production under the influence of the Antarctic
ozone hole. Icecolors ?90. Pages 159?186 in C. S. Weiler
and P. A. Penhale, editors. Ultraviolet radiation in Antarc-
tica: measurement and biological effects. American Geo-
physical Union, Washington, D.C., USA.
Quesada, A., J.-L. Mouget, and W. F. Vincent. 1995. Growth
of Antarctic cyanobacteria under ultraviolet radiation: UVA
counteracts UVB inhibition. Journal of Phycology 31:242?
248.
Roy, S. 2000. Strategies for the minimisation of UV-induced
damage. Pages 177?205 in S. de Mora, S. Demers, and M.
Vernet, editors. The effects of UV radiation in the marine
environment. Cambridge University Press, New York, New
York, USA.
Rundel, R. D. 1983. Action spectra and estimation of bio-
logically effective UV radiation. Physiologia Plantarum 58:
360?366.
SAS Institute. 1996. SAS for Macintosh, release 6.12. SAS
Institute, Cary, North Carolina, USA.
Schindler, D. W., P. J. Curtis, B. R. Parker, and M. P. Stainton.
1996. Consequences of climate warming and lake acidi-
fication for UV-B penetration in North American boreal
lakes. Nature 379:705?708.
Scully, N. M., and D. R. S. Lean. 1994. The attenuation of
ultraviolet radiation in temperate lakes. Archiv fu?r Hydro-
biologie Beihefte Ergebnisse der Limnologie 43:135?144.
Setlow, R. B. 1974. The wavelengths in sunlight effective in
producing skin cancer: a theoretical analysis. Proceedings
of the National Academy of Sciences (USA) 71:3363?
3366.
Shindell, D. T., D. Rind, and P. Lonergan. 1998. Increased
polar stratospheric ozone losses and delayed eventual re-
covery owing to increasing greenhouse-gas concentrations.
Nature 392:589?592.
Smith, K. C., and E. R. Macagno. 1990. UV photoreceptors
in the compound eye of Daphnia magna (Crustacea, Bran-
chiopoda). A fourth spectral class in single ommatidia.
Journal of Comparative Physiology A 166:597?606.
Smith, R. C., K. S. Baker, O. Holm-Hansen, and R. Olson.
1980. Photoinhibition of photosynthesis in natural waters.
Photochemistry and Photobiology 31:585?592.
December 2001 1857UVR SPECTRAL COMPOSITION AND IMPACT
Smith, R. C., B. B. Pre?zelin, K. S. Baker, R. R. Bidigare, N.
P. Boucher, T. Coley, D. Karentz, S. MacIntyre, H. A. Ma-
tlick, D. Menzies, M. Ondrusek, Z. Wan, and K. J. Waters.
1992. Ozone depletion: ultraviolet radiation and phyto-
plankton biology in Antarctic waters. Science 255:952?
959.
Sommaruga, R., I. Obernosterer, G. J. Herndl, and R. Psenner.
1997. Inhibitory effect of solar radiation on thymidine and
leucine incorporation by freshwater and marine bacterio-
plankton. Applied and Environmental Microbiology 63:
4178?4184.
Stich, H. B., and W. Lampert. 1984. Growth and reproduction
of migrating and nonmigrating Daphnia species under sim-
ulated food and temperature conditions of diurnal vertical
migration. Oecologia 61:192?196.
Stolarski, R., R. Bojkov, L. Bishop, S. Zerefos, J. Staehelin,
and J. Zawodny. 1992. Measured trends in stratospheric
ozone. Science 256:342?349.
Storz, U. C., and R. J. Paul. 1998. Phototaxis in water fleas
(Daphnia magna) is differently influenced by visible and
UV light. Journal of Comparative Physiology A 183:709?
717.
Sutherland, B. M. 1981. Photoreactivation. BioScience 31:
439?444.
Thorpe, J. E. 1977. Morphology, physiology, behavior, and
ecology of Perca fluviatilis L. and P. flavescens Mitchell.
Journal of the Fisheries Research Board of Canada 34:
1504?1514.
Vernet, M., A. Neori, and F. T. Haxo. 1989. Spectral prop-
erties and photosynthetic action in red-tide populations of
Prorocentrum micans and Gonyaulax polyedra. Marine Bi-
ology 103:365?371.
Williamson, C. E. 1993. Linking predation risk models with
behavioral mechanisms: identifying population bottle-
necks. Ecology 74:320?331.
Williamson, C. E., B. R. Hargreaves, P. S. Orr, and P. A.
Lovera. 1999. Does UV play a role in changes in predation
and zooplankton community structure in acidified lakes?
Limnology and Oceanography 44:774?783.
Williamson, C. E., S. L. Metzgar, P. A. Lovera, and R. E.
Moeller. 1997. Solar ultraviolet radiation and the spawning
habitat of yellow perch, Perca flavescens. Ecological Ap-
plications 7:1017?1023.
Williamson, C. E., O. G. Olson, S. E. Lott, N. D. Walker, D.
R. Engstrom, and B. R. Hargreaves. 2001. Ultraviolet ra-
diation and zooplankton community structure following de-
glaciation in Glacier Bay, Alaska. Ecology 82:1748?1760.
Williamson, C. E., R. S. Stemberger, D. P. Morris, T. M. Frost,
and S. G. Paulsen. 1996. Ultraviolet radiation in North
American lakes: attenuation estimates from DOC mea-
surements and implications for plankton communities.
Limnology and Oceanography 41:1024?1034.
Williamson, C. E., and H. E. Zagarese, editors. 1994. Impact
of UV-B radiation on pelagic freshwater ecosystems. Ar-
chiv fu?r Hydrobiologie Beihefte Ergebnisse der Limnologie
43:1?226.
Williamson, C. E., H. E. Zagarese, P. C. Schulze, B. R. Har-
greaves, and J. Seva. 1994. The impact of short-term ex-
posure to UV-B radiation on zooplankton communities in
north temperate lakes. Journal of Plankton Research 16:
205?218.
Worrest, R. C. 1982. Review of literature concerning the
impact of UV-B radiation upon marine organisms. Pages
429?457 in J. Calkins, editor. The role of solar ultraviolet
radiation in marine ecosystems. Plenum, New York, New
York, USA.
Yan, N. D., W. Keller, N. M. Scully, D. R. S. Lean, and P. J.
Dillon. 1996. Increased UV-B penetration in a lake owing
to drought-induced acidification. Nature 381:141?143.
Young, A. R., L. O. Bjo?rn, J. Moan, and W. Nultsch. 1993.
Environmental UV photobiology. Plenum, New York, New
York, USA.
Zagarese, H. E., and C. E. Williamson. 1994. Modeling the
impacts of UV-B radiation on ecological interactions in
freshwater and marine ecosystems. Pages 315?328 in R.
H. Biggs and M. E. B. Joyner, editors. Stratospheric ozone
depletion/UV-B radiation in the biosphere. Springer-Ver-
lag, Berlin, Germany.
Zagarese, H. E., C. E. Williamson, M. Mislevets, and P. Orr.
1994. The vulnerability of Daphnia to UV-B radiation in
the Northeastern United States. Archiv fu?r Hydrobiologie
Beihefte Ergebnisse der Limnologie 43:207?216.