Seasonal variation of the water quality of rivers and streams of eastern Mediterranean

Accepted 21 January 2002 Copyright © EEF ISSN 1399-1183 The use of biological methods for monitoring (biomonitoring) the ecological quality of running waters is already a widespread approach in Europe. It provides information on water quality and ecology both before and during sampling, in contrast to chemical approaches which only characterize the water system during sampling (Metcalfe 1989). However, a combination of both approaches, physicochemical and biotic, has been shown to be the most appropriate method for pollution monitoring (Hellawell 1986, Metcalfe 1989, Mason 1991). Biomonitoring includes both sublethal changes at the cellular or tissue level (Vukmirovic et al. 1994, Cajaraville et al. 1995) and changes in community (Metcalfe 1989, Graca et al. 1989, Castella et al. 1995). The use of changes in community structure for monitoring freshwater pollution, commonly involves benthic macroinvertebrates, because this group is considered as the most appropriate biological indicator of water quality in EU countries (Metcalfe 1989) including Greece (Anagnostopoulou 1992, Anagnostopoulou et al. 1994). Benthic macroinvertebrates are the most appropriate biotic indicators for the following reasons: 1) These organisms are relatively sedentary and are therefore representative of local conditions. 2) Macroinvertebrate communities are very heterogeneous, consisting of representatives of several phyla. The probability that at least some of these organisms will react to a particular change of environmental conditions, is therefore high (De Pauw and Vanhooren 1983, Hellawell 1986, Metcalfe 1989, Mason 1991). Other groups of organisms (fish, phytoplakton, etc) possess some, but not all, of these imSeasonal variation of the water quality of rivers and streams of eastern Mediterranean

The use of biological methods for monitoring (biomonitoring) the ecological quality of running waters is already a widespread approach in Europe.It provides information on water quality and ecology both before and during sampling, in contrast to chemical approaches which only characterize the water system during sampling (Metcalfe 1989).However, a combination of both approaches, physicochemical and biotic, has been shown to be the most appropriate method for pollution monitoring (Hellawell 1986, Metcalfe 1989, Mason 1991).
Biomonitoring includes both sublethal changes at the cellular or tissue level (Vukmirovic et al. 1994, Cajaraville et al. 1995) and changes in community (Metcalfe 1989, Graca et al. 1989, Castella et al. 1995).The use of changes in community structure for monitoring freshwater pollu-tion, commonly involves benthic macroinvertebrates, because this group is considered as the most appropriate biological indicator of water quality in EU countries (Metcalfe 1989) including Greece (Anagnostopoulou 1992, Anagnostopoulou et al. 1994).Benthic macroinvertebrates are the most appropriate biotic indicators for the following reasons: 1) These organisms are relatively sedentary and are therefore representative of local conditions.2) Macroinvertebrate communities are very heterogeneous, consisting of representatives of several phyla.The probability that at least some of these organisms will react to a particular change of environmental conditions, is therefore high (De Pauw and Vanhooren 1983, Hellawell 1986, Metcalfe 1989, Mason 1991).Other groups of organisms (fish, phytoplakton, etc) possess some, but not all, of these im-

M. Lazaridou-Dimitriadou
Lazaridou-Dimitriadou, M. 2002.Seasonal variation of the water quality of rivers and streams of eastern  Biotic and abiotic data on undisturbed or moderately disturbed lotic sites from a number of studies carried out in northern Greece showed that large rivers differ from small rivers, streams or creeks in terms of diversity, dominant groups and the kind of taxa (concerning the sensitivity of the taxa according to Biological Monitoring Working Party (BMWP) biotic scores).This is mainly due to the differences in their physical characteristics.Correlation of the environmental variables using MDA (multiple discriminant analysis) showed that the chief differentiating factors among the above water bodies are substrate, total suspended solids (TSS), conductivity, slope and temperature.Additionally, there is no clear phenological seasonality in the majority of the dominant benthic macroinvertebrate groups when undisturbed or moderately disturbed sites of mountainous creeks and small rivers are examined.By contrast, in downstream sites of long rivers, seasonality characterizes the dominant benthic macroinvertebrate groups, as it does for other Mediterranean animals.
M. Dept of Zoology,School of Biology,Fac. of Science,Aristotle Univ.,Thessaloniki,Greece.portant attributes.3) Macroinvertebrates are differentially sensitive to pollutants of various types, and react to them quickly; also, their communities are capable of a gradient response to a broad spectrum of kinds and degrees of stress.4) Their life spans are long enough to provide a record of environmental quality.5) Macroinvertebrates are ubiquitous, abundant and relatively easy to collect.Furthermore, their indentificaton and enumeration is not as tedious and difficult as that of microorganisms and plankton.
Greece is one of the EU countries that have no specialized institution to undertake the regular monitoring of running waters.During the last seven years, two large rivers (Aliakmon and Axios) and five small ones (Aggitis, Maurolakkas, Basdekis, Kokkinolakas, and Asprolakas) have been studied in the N. Greece in order to assess an appropriate method for biomonitoring the ecological quality of Hellenic freshwater ecosystems as this is determined by the EU Proposal Directive [Commission of the European Communities COM (1998) 76 final] (Anagnostopoulou 1992, Anagnostopoulou et al. 1994, Copeland et al. 1997, Ford et al. 1998, Langrick et al. 1998, Artemiadou et al. 1999, Drouin et al. 1999, Jennings et al. 1999, Yfantis et al. 1999).The above studies involve the comparative application of several European biotic indices, scores and multivariate statistical methods, to establish a Hellenic biotic index, which would be applied to the special hydrological and climatological conditions in Greece (high temperatures during summer, strikingly variable discharge throughout the year) because the applied European biotic indices and scores do not totally match (Lazaridou-Dimitriadou et al. 2000).
In the present study, various undisturbed sites from the above studies were examined at monthly (Axios and Aliakmon rivers) or seasonal basis (Olympias and Skouries streams) in order to find out the factors affecting macrobenthic community structure in rivers and streams and to investigate seasonal patterns of macroinvertebrate abundance and diversity.For the above comparisons the seasonal physicochemical and benthic macroinvertebrate community structure and abundance were analysed from undisturbed or moderately disturbed sites, one at the upper reaches and the other at the downstream stretch, in two rivers (Aliakmon and Axios) and two streams (Piavitsa-Asprolakkas, Kipouristras) of N. Greece.

Study area
Available data were limited in this region as is mentioned above.Data available from past studies were used in this review (Ford et al. 1998, Langrick et al. 1998, Bobori and Mourelatos 1999, Drouin et al. 1999, Jenning et al. 1999, Yfantis et al. 1999, Lazaridou-Dimtiriadou et al. 2000).

Skouries and Olympias streams
Skouries and Olympias areas are located on the east coast of Chalkidiki in northern Greece.All streams originate from the Stratonikon mountain range and flow through a mountainous landscape, except for some flat areas close to Olympias town.
Concerning the Olympias area, Kipouristras Creek (which has two sampling stations) flows from south to north in a semi-mountainous area before its confluence with Mavrolakkas River (three sampling stations).The latter follows a west-to-east course and then discharges into the gulf of Strimonikos, ca 1 km north of Olympias town.Within the area of Olympias, agricultural activities are relatively developed and concentrated around the flat regions close to Olympias town (Table 1).These areas also support a number of livestock farms with semi-wild pigs and goats.The main human activities in the area are related to ore mining.These include both old and currently active mine shafts, a mill plant for mineral enrichment and a backfilling plant, which are currently not fully operating, ore (mainly arsenopyrite) stockpiles, waste rock deposits (Platia Fire), an old tailings pond and four settling ponds for the precipitation of metals contained in ground water flowing from the old adits.
The Skouries area, is a mountainous region, is largely covered by forest.Its north side is adjacent to an area protected by "Natura 2000" EU Directive, located in between Olympias and Skouries.At present the main human activities within this study area concern woodcutting, and stock-raising (Table 1).Asprolakkas stream is the main stream that runs through Skouries area and derives from the confluence of Kerasia and Piavitsa creeks.

Aliakmon river
The springs of the river Aliakmon is located in Grammos and Vernon Mountains, in northwestern Greece.It discharges into the gulf of Thermaikos.Upstream sites, located along the main course of the river before the artificial lake of Polyphyto, do not receive significant amounts of pollutants apart from Kostarazi (sewage wastes) and Greveniotis (sewage and industrial wastes).Downstream sites, located in tributaries of the Aliakmon (the rivers Moglenitsas, Edesseos and the drainage channel, Canal 66) are subject to agricultural (pesticides and fertilizers), urban (sewage and detergents) and industrial (mainly fruit and vegetable canneries) sources of pollution (Anagnostopoulou 1992, Kouimtzis et al. 1994).The physical characteristics of the studied sites of this river are shown in Table 2.

Axios river
The river Axios originates in the Sar mountains of the Former Yugoslav Republic of Macedonia (FYROM) and discharges into the Thermaikos Gulf to the south-west of the city of Thessaloniki in northern Greece.Of its total length of 320 km only the last 80 km is within Greek territory and the study area is thus confined to the lower reaches of the river.Catchment land-use is predominantly agricultural and diffuse pollution as a result of soil erosion and pesticide and fertiliser run-off may be expected.Local settlements and industries are small, but pollution loads may also arise due to point source domestic wastewaters (including sewage and detergents), industrial effluents (including textile, slaughterhouse and dairy wastes), and polluted tributaries (carrying agricultural drainage waters).
The physical characteristics of the studied sites of this river are shown in Table 2.

Sampling of macroinvertebrates and physicochemical parameters
One sample of benthic macroinvertebrates was taken into consideration from one site at the upper reaches and one site at the downstream stretch of each of the rivers or streams mentioned above.This is clearly shown by the difference in the substrate composition between the two sites at least in the long rivers Aliakmon and Axios (Table 3).Details of the sampling procedure, the treatment of samples, and the measurement of the different physicochemical parameters are given in a previous paper (Lazaridou-Dimitriadou et al. 2000).

Statistical analyses
In order to compare the biotic parameters in the above mentioned lotic systems (rivers and streams), seasonal data (physicochemical data) from an upstream site and a downstream stretch were examined and analysed as well (Tables 5-9).Several criteria had to be met for inclusion of these undisturbed or moderately disturbed sites into the dataset (Jennings et al. 1999): 1) Taxonomic resolution had to be comparable throughout the dataset.This forced the recombining of several species to genus or family level.2) Each site had to be sampled in at least two seasons (spring and autumn) to lessen the chance of seasonal taxa being missed.3) Sites had to be relatively unstressed.
Adhering to these criteria we produced a dataset of 24 sites (5 from creeks or streams in Olympias area, 8 from creeks or streams in Skouries area, 9 in the Aliakmon and two in the Moglenitsas river which merges with Aliakmon) from a total of 268 studied sites.Only spring and autumn data were used.A list of 131 taxa was compiled, identified mostly to genus level.Exceptions were Oligochaetes (to class) and Diptera (to family).Sites and species classified by the clustering FUZZY technique (Equihua 1990), based on fuzzy sets, were used in Multiple Discriminant Analysis (MDA) relating the FUZZY site groupings to their environmental variables.MDA was implemented through the discriminant function of the computer program SPSS (Klecka 1975).

Seasonality
In order to examine whether seasonality affects the phenology of benthic macroinvertebrate community structure the abundance of each benthic taxon (sampled with the three min kick/sweep method) was monthly or seasonally followed at two representative sites, one at the upper reaches and the other in the downstream stretch in each of the mentioned above rivers or streams.

Correlation of site groupings with environmental parameters
MDA was used to attempt to predict the FUZZY groups of sites using environmental data only.The examination of the standardised canonical discriminant functions gave an indication of the importance of the various environmental parameters contributions to the prediction (Table 4).
In the first discriminant function the five most important variables (substrate, total suspended solids (TSS), conductivity, slope and temperature) seemed to be related to the longitudinal positioning of a site on a water course.

Seasonality
In all the Greek rivers and creeks, mainly because of reduced river vegetation, total suspended solids (which according to MDA is one of the most important discriminating environmental variables) increase (Bobori and Mourelatos 1999) after a rainy season or after snow melting or both (Tables 5 and 6).This causes differences in the makeup of biocommunities (Tables 5 and 6).High discharge (in January) are favoured by the sensitive macroinvertebtate groups whereas tolerant macroinvertebrate groups appear in the beginning of spring (Table 10A, B).Nutrients exceed the EU limits in the two studied rivers only during summer (Tables 8 and 9).
In Aliakmon river Plecoptera mainly appear in January, Trichoptera mainly in July-August and then in October.Ephemeroptera, although some genera mainly appear from October to February and others in June (Table 10A), as well as Oligochaeta and Chironomidae, appear all the year round (Fig. 1a, Table 10A).In Axios river there are almost no Plecoptera since Axios represents the last 70 km of its length in the Greek territory.As to the other groups some Trichoptera appear mainly in June, Ephmeroptera appear from June to November, Oligochaeta from January to May, Chironomidae from June to February showing one peak in summer and one in winter (Fig. 1b, Table 10B).On the other hand, in Olympiada and Skouries creeks, all the above groups appear throughout the year, except for Ephemeroptera that peak in June and Oligochaeta that peak in August-November (Table 10C, D).
Comparing the sum of the number of families seasonally, over the months January, March, June, August and November, it is clear that it declines ca 70% from the small (Olympias and Skouries areas) to the large rivers (Aliakmon) and up to 85% from the small to the plain parts of long rivers (Axios river) (Table 10).Additionally, comparing the scores of sensitive or moderately sensitive taxa, with BMWP scores (Chesters 1980) of 10 down to 5, seasonally it is noticed that 80% more sensitive taxa are found in small (as Skouries and Olympias) and large rivers (as Aliakmon) than in the plain parts of long rivers (e.g.Axios river).

Discussion
Concerning the studied sites, the results suggest that the substrate and TSS were found to be the most important in the studied sites in determining community structure.Conductivity, slope and temperature were also found to be highly influential factors.The influence of substrate on macroinvertebrate communities, found as a differentiating factor by MDA, has long been known as such (Hynes 1970).In Greek rivers, discharge is much higher during winter and the beginning of spring whereas conductivity declines.The latter increases in summer, mainly in June, causing a deterioration of biocommunities in summer (June-August).The results of this study vary from others in the strength of chemical parameters as influencing fac-tors on community structure.Wright (1995) found alkalinity to be the most important factor by MDA, whilst Marchant et al. (1994) found a strong pH and conductivity gradient.Such strong gradients are not apparent in this study apart from conductivity.This is likely to be because such parameters are of more use in detecting differences between, rather than along rivers.Wright's (1995) and Marchant's et al. (1994) studies investigated 41 and 9 rivers respectively, whilst this study investigates only 4 systems (Olympias, Skouries, Aliakmon and Moglenitsas).Therefore factors relating to differences between rivers were of less importance in this study.
Concerning the distribution of various taxa along the environmental gradients it is clear that Trichoptera largely show a preference for more diverse, coarser substrates, rather than depositional ones.The Ephemeroptera show a more general distribution, with different taxa showing preferences for both streams and larger river sites.The Potamanthidae were found only in the larger Aliakmon sites, where they live on the sand and gravel.The prosobranch families (especially the Hydrobiidae) were largely confined to turbulent stream sites due to their higher dissolved oxygen requirements (e.g.Skouries and Olympias).
Conversely, the pulmonates (Physidae) were found to flourish mainly in the Aliakmon and less in the Axios sites, due to their ability to exploit less turbulent microhabitats.
Concerning seasonality, the study of the two representative sites in each Greek lotic system showed that streams and creeks must be regarded separately from large and long rivers according to the MDA analysis, because these systems differ mainly in their physical characteristics and less on the strength of their chemical characteristics.This can perhaps be explained by the fact that the studied small rivers, here in Greece, run mainly in woody areas and their physical characteristics do not change much along them, so biocommunities do not have to accomodate to striking differences.Whereas in the plain parts of long rivers, as is the case for Axios river, different land uses along them cause differences in the physicochemical parameters.So, the different benthic macroinvertebrate groups flourish at different times of the year, taking advantage of the most favourable conditions for them to fulfill their cycles.Consequently, the evolution of the biocommunity structure in   the small rivers, streams or creeks is free of distinct seasonality in phenology for 50% of dominant benthic macroinvertebrate groups.In the plain parts of large rivers, however, seasonality characterizes all the groups as happens with other animal groups in the Mediterranean region (Argyropoulou et al. 1993, 1994, Stamou et al. 1993).Accordingly, in Axios Ephemeroptera and Chironomidae peak at the same time (June-November) whereas Oligochaeta, that occupy almost the same habitat with Chironomidae, peak from November to June when water quality is poor mainly due to TSS.

Fig. 1 .
Fig. 1.Monthly fluctuations of the abundance of the benthic macroinvertebrate groups in two sites in Aliakmon (a) and Axios (b) river.

Table 1 .
Physical characteristics and sources of pollution of the sampling stations of Kipouristras stream at Olympias area (

Table 2 .
Physical characteristics and sources of pollution of the sampling stations of Aliakmon (Grevena and Niseli) and Axios rivers (Euzoni and Anatoliko).
*This distance is calculated from the borders with FYROM.

Table 4 .
Standardized canonical discriminant function coefficients of environmental variables.Only the 10 most influencing parameters are shown and ranked.

Table 9 .
Nutrients at Olympias and Skouries areas (bold values: values exceeding the EU limits).