3.1.1. Differences in scales and responses

Regulation of the water level does not introduce any new environmental factors affecting the lake littoral; all the factors present in unregulated lakes also exist in regulated lakes (I). There are, however, significant changes in their amplitude and timing, which usually results in a more vulnerable environment for aquatic macrophytes (Rørslett 1988a, I). My results, which were obtained by comparing two different lakes, support the comparability of the two lakes and are largely applicable to other lakes in Fennoscandia. As pointed out in the papers I and II, the environmental factors can be divided into three groups: lake-, shore- and site-specific factors (Fig. 2).

Lake-specific factors are similarly distributed within each lake; water level fluctuation and the related ice factors are a typical example (Figs. 2-3). In general, water quality is also lake-specific, but it may vary widely in a large lake divided into sub-basins. The elevation of water level at the beginning of regulation is also a lake specific background factor, which affects the stability of the shoreline. Shore-specific factors consist of geomorphological factors, which vary depending on the exposure within the lake. These mainly horizontally oriented factors are driven by geomorphological processes, which are only partly dependent on water level regulation. Site-specific factors are highly dependent on bottom quality, which correlates with geomorphological factors, exposure and depth.

My results showed that the lake-specific factors related to water level fluctuation can be divided into several sub-factors. Water level fluctuation during the open water period divides the littoral into different water level duration (d_w) zones (I). These zones called supra-, eu- and sublittoral explains largely the living areas of different life-forms. The supra- and eulittoral represents the living range of emergent plants, especially sedges (Carex spp.) (II). This vertical gradient is well known (Pearsall 1920, Lillieroth 1938, Vaarama 1938, Maristo 1941, Segal 1971, Eurola 1965, Andersson 1973, Hutchinson 1975, Spence 1982, Coops et al. 1996), but rarely described as duration zones (but see Pieczynska 1972, Mark & Johnson 1985, Blanch et al. 1999). The duration zone approach was used to estimate the expected living range of sedges with a relatively good fit (Hellsten et al. 1997, Riihimäki & Hellsten 1997, Nykänen 1998, Hellsten 2000). Real aquatic macrophytes (hydrophytes) always prefer the zone (sublittoral) below the median water level of the open water period (den Hartog & Segal 1964).

It seems that at least Carex rostrata may benefit from the delayed flood in lake Ontojärvi (II). The high environmental tolerance of C. rostrata is presumably related to its two-year life cycle (Bernard 1976), and the species achieves good resistance against flooding with the aid of aeronchymal tissues (Walker & Wehrhahn 1970, Bernard 1973, Elveland 1984, Fagerstedt 1992, Huttunen et al. 1996, Nykänen 1988). On the other hand, it decreases in abundance if the water level is raised (Nilsson 1977, Sjöberg & Danell 1983) and it can also grow without a high flood (Weiher & Keddy 1995). The responses of different graminoids are variable; Galamagrostis-species benefit from irregular water level (Gill 1973, Nykänen 1998), whereas Molinia caerulea (L.) Moench is completely lacking due to its sensitivity to prolonged flooding (Kotilainen 1958, Wassen 1966, Nilsson 1981b, Fraisse et al. 1997).

Large helophytes (Equisetum fluviatile L., Phragmites australis) are also typical of the fluctuation zones, but due to their higher stems, they are able to survive at a deeper level (II). Their frequency has clearly declined in Lake Ontojärvi due to the marked erosion. E. fluviatile does not suffer from a high water level during dormancy (Pearce & Cordes 1988), but a high water level during the growing season causes a permanent decline of this species (Rintanen 1976, Anttonen-Heikkilä 1983, Toivonen & Nybom 1989, Wallsten & Forsgren 1989, van den Brink et al. 1995). The changed water level dynamics together with abundant erosion create difficult living conditions for E. fluviatile, causing a sudden decline in Equisetum stands.

The disappearance of Phragmites australis is also well known in regulated water courses (Nilsson 1978, Granberg & Hakkari 1980, Hellsten & Joronen 1986, II). It benefits from the low water levels during the early summer (Coops & Van der Velde 1995), and stabilisation of the water level will cause a decline in its distribution (Coops et al. 1996). It also suffers from organic rich bottom quality (e.g. Ostendorp 1989, Weisner 1991, Coops et al. 1996, Clevering 1999). From the water level fluctuation and bottom quality point of view, there should be luxuriant growth of Phragmites in my research lakes. On the other hand, it does not tolerate high exposure (Coops et al. 1991, Coops et al. 1994) and is also sensitive to ice erosion (e.g. Luther 1951a). The permanent plots (III) and planting experiments in Lake Ontojärvi (IV) showed that Phragmites, being a good competitor, is able to retain its growing areas, but unable to expand its living range. Obviously, bottom instability combined with continuous ice and wave erosion restricts new occurrences.

My study also showed the essential need to differentiate between observed depth (moving Lagrangian co-ordinate) and depth calculated from the open-water mean (fixed Eularian co-ordinate) as proposed by Rørslett (1984, 1985a, 1987a, b, 1988a). Especially in regulated lakes with fluctuating water levels, the use of an exact datum level is a basic assumption in calculations.

The effect of ice expanding to the bottom can be divided into two strikingly different factors: the frozen bottom zone is situated in the upper part of the littoral and characterised by frozen sediment up to 40 cm in thickness (I). This is typical of minerogenic shores, where the lowest level of the frozen bottom zone can be predicted by the date when the lowering ice touches the bottom. Most of the peat bottoms and bottoms with high organic contents remain non-frozen, as also noted by Palomäki & Koskenniemi (1993) and Koskenniemi (1994). This date in the first week of February is relevant in the northern part of Finland, but it also seems to be applicable in the western part of Finland (Hellsten et al. 2000). Comparable values can be noticed from northern Swedish reservoirs and rivers too (Nilsson 1981a, Erixon 1979,1981, Renman 1989,1993, I). Functionally, the same phenomenon has been described as an ice scouring function by Rørslett (1984, 1985a, 1987b, 1988a). The zone below the frozen zone is only affected by the mechanical pressure of ice and can be called an ice penetration zone. In this zone, the ice merely presses down on the bottom, but the surface of sediment is not frozen. These two zones can be described in terms of the duration of the frozen (d_f) and ice penetration (d_p) zones (I).

The responses of the species can be divided into three basic types. Helophytes, which are bounded to the upper zone, and small isoetids (Ranunculus reptans, Eleocharis acicularis) are relatively tolerant against bottom freezing and prefer the frozen zone (II). The second group consists of large isoetids, such as Lobelia dortmanna and Isoetes echinospora Durieu, which avoid the frozen zone but are able to grow in the ice penetration zone (II). The third group is characterised by Isoetes lacustris and some large nymphaeids (Potamogeton natans L., Nuphar lutea (L.) Sibth. & Sm), which avoid the ice pressure zone (II).

Small isoetids are well adapted to the vulnerable environment present in the upper zone of the lake and are therefore much more common in Lake Ontojärvi than Lake Lentua (II). Due to their flexible growth strategies, they can spread rapidly over the entire littoral (e.g. Rørslett 1989). It has been even stated that Ranunculus reptans benefits from being broken up by ice into small pieces capable to spread vegetatively (Renman 1989). These two species are found in extreme environments with water level fluctuation up to 20 meters (Nilsson 1978, 1981a, Granberg & Hakkari 1980, Renman 1986, Wilcox & Meeker 1991, Hill et al. 1998). However, Rørslett (1989) noted that R. reptans was lacking in a lake with 14.4 meter regulation, but was able to return to the bare exposed littoral of the regulated reservoir Meltingen (21 meters of regulation) after the addition of fertilisers (Rørslett & Johansen 1996).

Large isoetids clearly suffer from the effects of ice (II). The most sensitive species, Isoetes lacustris, even avoided the zone of penetrating ice due to its growth form, which is characterised by hard leaves (Aulio 1985, Rørslett & Brettum 1989). Its sensitivity to ice has been described in several studies (Quennerstedt 1958, Nilsson 1978, Renman 1989, Rørslett 1984, 1985a, b, 1987b, c, 1988b, 1989, Rintanen 1996). On the other hand, Isoetes echinospora with its flexible habitus with bendable leaves is able to survive in Lake Ontojärvi very well (II). It is notably ice-resistant (Luther 1951a, Nilsson 1978, Rørslett & Brettum 1989) and obviously benefits from the disappearance of I. lacustris in the lower littoral. On the other hand, the absence of helophytes in the upper littoral zone may stimulate its growth in Lake Ontojärvi.

The distribution of Lobelia dortmanna is very limited in Lake Ontojärvi (II). Its decline has been demonstrated in several regulated lakes (Rørslett 1985ab, 1987c, 1989). It usually prefers minerogenic bottoms with low nutrient contents (e.g. Kurimo 1970, Szmeja 1987, Toivonen & Huttunen 1995) due to its large root system specialised to CO₂ intake (Wium-Andersen 1971). On the other hand, it cannot escape to deeper area by increasing the amount of chlorophyll (Kansanen & Niemi 1974). Minerogenic shore areas are also easily frozen (Palomäki & Koskenniemi 1993, I), and the damage to the root system is therefore obvious.

Another group sensitive to ice consists of large nymphaeids, such as Potamogeton natans and Nuphar lutea (Luther 1951b, II). P. natans prefers soft bottoms, which are usually not frozen, and it can therefore be found even in the littoral of Lake Ontojärvi and regulated reservoirs (Koskenniemi 1987, Erixon 1981). The distribution of N. lutea was quite similar to that of P. natans; a soft unfrozen bottom enabled it to occupy the ice pressure zone. In general, the decrease in the frequency of N. lutea is very clear (Nilsson 1978, Erixon 1979).

In addition to the water level fluctuation and ice extension, the attenuation of light forms a depth related pattern which restricts the lowest level of aquatic macrophytes (I, II). In my study, it is described by the intensity of red light (L_D). A typical example of light limited distribution is the distribution of Isoetes lacustris, which is almost lacking in Lake Ontojärvi (II). The relationship between red light and the lowest depth limit of I. lacustris was compatible with the 4-6 % observation of Eloranta & Marja-aho (1982). The ecological niche of I. lacustris between the ice disturbance and the lack of light has been described in detail in several studies of Rørslett (1984, 1985a, b, 1987b, c, 1988b, 1989).

All of these factors are clearly related to elevation (z) or depth (D), which means that their values remain constant at same vertical level of the littoral within a given lake (Fig. 2). These factors are heavily affected by the water level fluctuation. Their values can be easily used to evaluate the effects of different water level regulation practices (V).

The shore-specific factors consist of geomorphological factors (slope S, shape C) and effective fetch (F_e), which fluctuate quite randomly within a given lake (I, Fig. 3). These horizontal shore-specific factors are mainly driven by geomorphological processes, which are only partly dependent on water level regulation. It should been noted that the changes in water level greatly affect the significance of fetch, because the stability of bottom decreases if the water level is raised or lowered (Sundborg 1977, Granberg & Hakkari 1980, Nilsson 1981a, Hellsten & Alasaarela 1984, Newbury & McCullogh 1984, Hellsten & Joronen 1986, Mark 1987, Mark & Kirk 1987). The rise in water level, as seen in Lake Ontojärvi, launches abundant erosion, and the sandy shores of Lake Ontojärvi were much steeper compared to Lake Lentua, which shows the unstable nature of shores (I). On the other hand, the lowering of water level starts the increase of aquatic plants, as shown in Lake Oulujärvi (Nykänen 1998).

The response of macrophyte species to exposure is flexible. Aquatic macrophytes can partly escape into deeper water, while some species can arise their position to a higher level (Keddy 1982, Spence 1982). Exposure affects the distribution of helophytes and nymphaeids by mechanical force and indirectly via changes in bottom quality (e.g. Keddy 1982). My oligotrophic lakes showed no significant relationship between their presence and the exposure factors, but their presence is, however, quite clearly related to the bottom substrate (II). The diversity of macrophytes was slightly higher on sheltered shores (III, Nilsson 1981a, Keddy 1983). On the other hand, R­ørslett (1987c) and Nilsson & Keddy (1988) were unable to find any significant effect of exposure on the stability of vegetation.

It should also be noted that exposure values calculated as effective fetch (F_e) and shape (C) correlate significantly with each other (I, Palomäki & Hellsten 1996). The use of shape according to Palomäki (1992) divides the shores into erosional minerogenic cape areas and accumulative soft bottom bay areas, providing a clear insight of these discrete environments.

The inclination of shoreline or slope is assumed to be a significant factor affecting the diversity of the littoral flora (Duarte & Kalff 1986). Slope affects the distribution of bottom substrate (e.g. Håkanson & Jansson 1983), but there was only a weak correlation between slope and sedimentation level in my research lakes (I). There was therefore no correlation between the different species and continuous slope, but there were slightly more species on the mildly sloping shores (II, III).

Site-specific factors are illustrated by bottom quality, which itself is related to slope (S), shape (C) and fetch (F_e), but most clearly to water depth (Fig. 3). The shores of Lake Ontojärvi were steeper than those of Lake Lentua, which affected the distribution of bottom types, sandy bottoms being more common in Lake Ontojärvi than in Lake Lentua (I). The softness of the bottom usually increases along with increasing depth and decreasing exposure. The sedimentation level only correlated with the slope and was not predicted by fetch or shape. The distribution of bottom quality is difficult to predict. The main environmental factors from the two lakes were used in a discriminant analysis to predict the bottom type distribution of the littoral. Although the statistical significance of the predicted types was quite low (r2 = 0.41), it showed the possibility to predict the bottom quality by environmental data. In addition to these factors, the general soil types (moraine, glaciofluvial deposits, peatland, cliffs) of the area and the quality of water affect the formation of bottom quality.

The effect of bottom quality on aquatic macrophytes is partly an artefact (II). Carex species are always found on peaty bottom, but they actually prefer sheltered, mildly sloping shores and themselves produce peat. On the other hand, nymphaeids mainly grow on soft bottoms, which exist in sheltered bays suitable for species with floating leaves sensitive to wave action. The same also applies to large isoetids, whose correlation with soft bottom is related to the fact that soft bottom is typical of deeper areas (I, II). My observations are largely concordant with those of Rørslett (1985b), who was not able to find any relationship between bottom quality and macrophytes. The importance of bottom substrate has been shown by several other researchers (Pearsall 1920, Spence 1982, Barko & Smart 1983, Duarte & Kalff 1986, Wisheu & Keddy 1989). The inconsistent conclusions are mainly due to the different points of view, but at least in my research, the lake bottom substrate is a consequence of exposure and depth rather than an independent environmental factor in itself.

Despite the different environmental regimes, several common vegetation types were found in a TWINSPAN analysis (II). Most of the types were present in both lakes, but due to the harsh environment in Lake Ontojärvi, their ecological niches were very limited. This was especially true of ice-sensitive, large isoetids, such as Isoetes lacustris and Lobelia dortmanna. In general, the analysis indicated a similar structure in the vegetation in both lakes, although the different ecological regimes were reflected on the distribution of different types.

The disciminant analysis according the method of Moss et al. (1987) was used to predict the vegetation type distribution of the littoral (II). The analysis classified 50.1 % and 45.1 % of the vegetation types correctly in Lake Ontojärvi and Lake Lentua, respectively. Although the hit percentage was relatively low, it should be noted that the original TWINSPAN classification was very detailed, including 12 types in Lake Ontojärvi and 14 types in Lake Lentua. In general, the predicted types were ecologically close to the observed types, and a simplification of typology would have improved the results. It was concluded that this prediction method provides useful tool for analysing the effects of water level regulation.

3.1.2. Dynamics of the littoral vegetation

The littoral is always partly unstable, as waves and currents cause erosion in the upper zone, whereas eroded material accumulates in the lower, deeper zone of the littoral (I). Furthermore, the differences between hydrological years cause differences in the environmental conditions of the upper zone. The rise of water level at the beginning of regulation also brings about a specific factor causing unbalance in vegetation. Nilsson et al. (1997) demonstrated that the abundance of species tends to decrease in storage reservoirs 30-40 years after their construction. Obviously, the same tendency is also present in regulated lakes in northern Finland, which are rarely affected by eutrophication.

In my study, the stability of vegetation was investigated using a permanent plot design (III). A follow-up of four years showed that the mean dissimilarities in Lake Ontojärvi and Lake Lentua were 0.238 and 0.297, respectively. The difference between the lakes was not significant. The effect of increased ecological stress was visible in the littoral zone; the number of species was lower in Lake Ontojärvi than in Lake Lentua. The lesser diversity is consistent with the observations of Grime (1979) and Rørslett (1989).

In Lake Ontojärvi, the area that is almost permanently submersed (sublittoral) showed higher dissimilarity (D_m) and lower diversity (N_sm) values compared to Lake Lentua (III). On the other hand, dissimilarity in the eulittoral zone was higher in Lake Ontojärvi compared to Lake Lentua. The stability of the water level during the open water period explains the high stability of vegetation in Lake Ontojärvi compared to the Swedish reservoirs of fluctuating water level of open water period (Nilsson 1981a, Nilsson & Keddy 1988).

The vegetation in both lakes was well adapted to the disturbance caused by waves and penetrating ice. The diversity and D_m values were lower on the exposed shores in both lakes compared to the sheltered shores. This agrees with the observations of several researchers (Keddy 1982, 1983, Nilsson 1981a). The gently sloping shores carried higher diversity, as also noted by Duarte & Kalff (1986). The peaty bottoms were the most stable environments, as also noticed by Koskenniemi (1987) and Nilsson (1981a). On the other hand, the muddy bottoms were unstable in our research lakes; these bottoms are situated at the deepest area of littoral, where the mortality of individuals is also high (Rørslett 1985b).

Generally speaking, the vegetation in Lake Ontojärvi is equally stable as that in Lake Lentua. Both diversity and dissimilarity values were slightly higher in Lake Lentua. The vegetation in regulated Lake Ontojärvi is well adapted to the ecological disturbance caused by the fluctuating water level. This does not mean that the succession is at its final stage, because the abundance of species is probably increasing and the differences between different hydrological years cause some unforeseen changes in the ecological environment (cf. Nilsson et al. 1997).

3.1.3. General overview

In general, the environment in the littoral of a regulated lake is much more vulnerable than that of a lake in its natural state (I, II). The most visible effects of water level regulation were related to the raised water level, which yielded erosion of sandy shores at the beginning of the regulation (I). Another direct effect of regulation was the altered fluctuation of the water level, which led to bottom instability and increased the area of the frozen and ice penetration zones. The effect of ice penetration was also easy to recognise on the shores of Lake Ontojärvi, where the surface sediment was frozen to a greater depth and across wider areas than in Lake Lentua. Below the freezing zone, the ice merely pressed down on the sediment. The vertical reduction of light was estimated on the basis of water colour.

The littoral area acts as an ecocline between the terrestrial and aquatic environments and serves as the habitat of aquatic macrophytes. The composition of vegetation in the littoral is difficult to predict. Environmental factors along with biogeographical factors (e.g. Whittaker 1977), herbivory (e.g. Jutila 1997) and competition (e.g. Grime 1977) constitute the basic patterns that affect the general distribution of plant species. Water level regulation incorporates a set of pronounced changes, which, in turn, complicate the interactions in the littoral ecosystem (Rørslett 1989, I, II, III).