Effects of different levels of UV-B radiation on marine epilithic communities : a short-term microcosm study *

A great environmental problem recognised during recent years is ozone depletion and a following increase of penetrating ultraviolet-B radiation (UVB, 280-315 nm). It has been well documented that UV-B radiation is capable of causing considerable damage to living organisms, both in aquatic and terrestrial habitats (e.g. Nolan and Amanatidis, 1995; Lean, 1998). The most obvious consequence of the increasing UV-B radiation in aquatic environment is an often reported decrease in primary production (e.g. El Sayed, 1988; Häder et al., 1998). The ability of UV-B to produce free radicals of great oxidative power is well known (Halliwell and Gutteridge, 1989), and the effects induced by enhanced exposure to UV-B at the cellular level are therefore complex and often of considerable damage to living organisms (for review see Hessen et al., 1997). Even present day UV-B radiation induces stress both for phytoplankton and aquatic macroalgae (Häder, 1997; Häder et al., 1998). As pointed out by many authors, UV-B related responses among algae are individual and depend both on the physiological stage of the organism and environmental conditions (e.g. Xiong et al., 1996; Häder, 1997; Sundbäck et al., 1997). Various species of algae differ in their sensitivity to UV-B radiation. Taking this fact into SCI. MAR., 64 (4): 363-368 SCIENTIA MARINA 2000

account, changes in species composition depending on the UV-B doses can be expected (Cullen and Neale, 1994;Villafane et al., 1995;Helbling et al., 1996;Wängberg et al., 1996).Consequently, successions at the level of primary producers would undoubtedly affect higher trophic levels and thus cause changes in the entire aquatic ecosystem.
However, much more research is required in order to understand the role of UV-B within algal communities (Häder et al., 1998).Benthic and epiphytic algal communities in particular seem to be poorly understood in this sense.Only a few experimental studies with natural communities have been performed (Vinnebrooke and Leavitt,1996;Sundbäck et al., 1997;Odmark et al., 1998).Vinnebrooke and Leavitt (1996) are the only ones who have taken epilithic communities in alpine lakes under consideration as well.In this study we performed experiments in order to study dose-related responses to UV-B radiation within epilithic communities in brackish water.We hope that our study will contribute to a better understanding of this important part of aquatic ecosystems.

MATERIALS AND METHODS
The study was performed in June 1999 at the Department of Applied Science of Mid Sweden University in Härnösand, east coast of Sweden (62°25'39N, 17°23'17E).Five similar stones were sampled at the depth of 0.5 m.The coastal area along Härnösand is tideless with brackish conditions (3.4-5.6‰) and chlorophyll a content of 2 mg/m 3 during June (County Administration, Härnösand).All macroinvertebrates were carefully removed from the stones collected, after which the stones were placed in a 40 l glass enclosure with artificial flow and aeration.All experiments were carried out at 23°C.The light/dark cycle was 16 h/8 h with an irradiance of 19 W m -2 (400-700 nm).
Epilithic communities were daily exposed to 1 h, 3 h, 5 h or 7 h of UV radiation (280-315 nm), respectively.One stone served as control.The stones were irradiated by UV-A radiation (315-400 nm) of 1.02 W m -2 and UV-B radiation of 0.73 W m -2 at a distance of 0.35 m from the light source.UV-A and UV-B radiation were measured with an IL 1400A broad band Radiometer (International Light, Inc. Newburyport, MA) equipped with detectors for UV-A (W#6259, #12667) and UV-B (W #6272, #12618).Ultraviolet radiation was obtained from two 40 W sunlamps (FS40, Westinghouse Elec.Corp. Lamp Div., Bloomfield, NJ, USA).The ultraviolet radiation was filtered through cellulose acetate (CA) film (0.13 mm thickness) in order to remove shorter wavelength components not encountered in nature.The CA was preburnt for 48 h at a distance of 1 m from 4 UV lamps in order to minimize the change of the filter properties.
The experiments were repeated three times and carried out indoors in order to keep physical culture conditions constant.The duration of the study was set to 7 days in order to avoid possible later effects of nutrient depletion.Epilithic communities were studied before (initial stage, day 0) and after the experiments (day 7).On each occasion three samples from each stone were collected in order to identify algae species and analyze community structures.Epilithic species were identified with the aid of a wide range of literature, the most important being Pankow (1990).A rather traditional algal system was used for the classification of the taxa found (van den Hoek et al., 1995).The frequency of each species present in the fixed samples was determined according to a relative classification: 1 -occasional, 2 -rare, 3 -frequent, 4 -dominant.The reason for the use of relative abundance was its high applicability showed both for monitoring of successions and calculation of ecological indices (e.g.Kangas et al., 1993;Vinebrooke et al., 1996;Danilov and Ekelund, 2000).
Changes in species composition and community structure were estimated with the aid of two different approaches: multivariate method (cluster analysis) and ecological diversity indices (Shannon-Wiener's and Hurlbert's).Richness of each group was estimated with the aid of species number.
Diversity indices used: 1. Shannon-Wiener (Wiener, 1948;Shannon and Weaver, 1949): , n = the number of individuals in a sample from a population, n i = the number of individuals in a species i of a sample from a population 2. Hurlbert's (Hurlbert, 1971) PIE (probability of interspecific encounters): N = the number of individuals in a community, p i = the fraction of a sample of individuals belonging to species i.
In addition the number of species was used as a simple estimate of community structure.All statistical analyses were performed in the computer package Minitab 11.0.

RESULTS AND DISCUSSION
At the initial stage Chlorophyceae, Cyanophyceae and Bacillariophyceae were represented by two, one and eight unambiguously identified species, respectively (Table 1).The results obtained during all three repeated experiments did not show any differences.Navicula-species were considered as spp.due to the extreme difficulties in identifying all of them.However, insecure identification could lead to erroneous conclusions from statistical analyses.Therefore, Navicula was treated as a whole at the genus-level.Codiolum was considered not as a separate genus but as a stage of a filamentous green alga, supposingly Hormiscia neglecta.However, due to its physiological validity Codiolum-stage was considered by all statistical analyses performed.At the initial stage epilithic communities were dominated by Pseudendoclonium marinum and Bacillariophyceae.Cyanophyceae were under-represented both qualitatively and quantitatively.
After seven days two members of Chlorophyceae -Monoraphidium contortum and M. minutum -were the most abundant species in the control, while Codiolum-stage disappeared.No Cyanophyceae were found.All diatoms survived (except Cocconeis pediculus) but their abundance decreased.
In all treatments with UV-B Cyanophyceae became the most abundant group.All Cyanophyceae found were filamentous forms.Correlation between abundance of Cyanophyceae and the dose of UV-B radiation was positive for all three members (Table 1), although only slightly positive (R=0.26) for Pseudanabaena mucicola.This pattern agrees well with that detected for periphyton and epilithic communities in an alpine lake (Vinnebrooke and Leavitt, 1996).The dominance of filamentous forms of Cyanophyceae is also consistent with the results reported earlier (Vinnebrooke and Leavitt, 1996;Odmark et al., 1998).We can conclude that Cyanophyceae seem to outcompete other algae groups under enhanced UV-B radiation in epilithic communities as well as in benthos and periphyton.
Another phenomenon was appearance and increase in abundance (positive correlation, R=0.33) of Cladophora sericea (Chlorophyceae) in all UV-B treated epilithic communities compared to the control.Monoraphidium contortum and M. minutum, being dominant in the control, were suppressed by all UV-B doses studied.Most of the diatoms correlated negatively with the increasing EFFECTS OF UV-B ON MARINE EPILITHON (MICROCOSM) 365 UV-B doses (R from -0.22 to -0.84).One of the possible explanations for the decrease in diatom abundance and richness and the suppressing of Monoraphidium contortum and M. minutum could be their relatively small size.There is evidence that small cells are more vulnerable to UV-B radiation than large cells (Laurion and Warwick, 1998).It has been shown that small diatoms became more damaged by UV-B than larger cells (Karentz et al., 1991).Therefore at the enhanced UV-B level a successions towards large cells can be expected.This pattern has been detected for diatoms in periphyton (Bothwell et al., 1993).In the present study, however, no such shifts have been observed.The diatom species present at the initial stage decreased considerably in their abundance or disappeared entirely (Fig. 1).However, the treatment for 5 h led to a sudden increase in the richness (but not in the abundance!) of Bacillariophyceae with the following drop at the treatment for 7 h.We believe that this fact could be considered as an artefact because of the exclusively occasional detection (single exemplars, relative abundance 1) of the species.One possible reason for this could be some microscopic dif-ferences in the stone surfaces which enabled hiding of some exemplars of the diatoms.However, the negative correlation to UV-B radiation observed for all diatoms in the present study disagrees with patterns observed for benthic communities on sandy sediment where under UVinduced stress diatoms codominated with cyanobacteria (Odmark et al., 1998).The epilithic communities in the present study responded quickly to UV-B treatment.No lag phase was observed as reported for sandy and periphytic communities (Bothwell et al., 1993, Odmark et al., 1998).Because of the experiment conditions used (see above) we believe that the patterns obtained were not additionally affected by nutrient limitation (Cullen and Lesser, 1991).
Negative effects of UV-B radiation on photosynthesis of algae are well known (e.g.Nielsen and Ekelund 1998), although sensitivity differs among species (Neale et al., 1998).UV-B radiation has been demonstrated to suppress growth and cell division (e.g.Keller et al., 1997).There is not much data about the individual sensitivity of periphytic algae to UV-B stress.The reason could be that the performance of experiments could differ between researchers, for example in UV-B doses and the wave spectrum of the radiation used.Therefore, it seems impossible to explain the patterns obtained in the present study, based on the published data on the sensitivity of different species.However, we suppose that the individual sensitivity to UV-B radiation of the species studied plays a key role in the succession observed.
The ecological indices calculated show a drop in diversity after 3 h exposure compared both to 1 h, on the one hand, and to 5 h and 7 h, on the other (Table 2).Diversity after 1 h and 5 h exposure was higher than in the control.It is generally believed that changing diversity of an aquatic community can be a good estimate of how favourable environmental conditions are (for review see e.g.Steneck and Dethier, 1994).If the increase in diversity observed after 5 h exposure to UV-B radiation should be viewed as artefact (see above), a slight decrease in 366 R.A. DANILOV and N.G.A. EKELUND diversity can be related to UV-B doses.However, we conclude that the shift in species composition was much more spectacular than the shift in diversity.Although we performed a relatively short-term study, changes in the structure of epilithic communities induced by the UV-B dose could be detected.Sundbäck et al. (1997) did not obtain any structural changes in benthic communities during a short-term experiment, which allows us to suppose that epilithic communities could be more vulnerable to UV-B radiation and therefore respond quicker than those on sandy sediments.
The multivariate cluster analysis based on the presence-absence with abundance matrix showed two large clusters: one built by the initial stage and control, and another consisting of the communities exposed to UV-B (Fig. 2).Thereby the communities which received higher doses of UV-B (3 h, 5 h and 7 h) are more closely related to each other than to the community treated for 1 h only.A possible explanation of this fact could be the specific sensitivity of individual species to UV-B stress mentioned above.We interpret the clustering of the control community together with the initial community as good evidence that the differences in patterns observed are directly related to the varying UV-B treatment.

TABLE 1 .
-The list of epilithic species identified at the initial stage (I.s.) of the microcosm and on the seventh day after continuous treatment with UV-B radiation for 0, 1, 3, 5 and 7 hours/daily, respectively.The digits indicate relative abundance of each species present.Only correlation coefficients (R) of the absolute value exceeding 0.1 are shown.

TABLE 2 .
-Diversity indices calculated for epilithic communities at the initial stage (I.s.) of the microcosm and on the seventh day after continuous treatment with UV-B radiation for 1, 3, 5 or 7 h/day, respectively.