Implications of rapid changes in chlorophyll-a of plankton, epipelon, and epiphyton in a Pampean shallow lake: an interpretation in terms of a conceptual model

Biomass assessments of algae in wetlands usually include only the phytoplankton community without considering the contribution of other algal associations to total algal biomass. This omission prevents an accurate evaluation of the phytoplankton community as an integral part of the total ecosystem. In the present work, the biomass contributions (expressed as chlorophyll-a content per m2 of lake) of phytoplankton, epiphyton on both submerged and emergent macrophytes, and epipelon were measured in Lacombe Lake, Argentina, for the purpose of (1) establishing the relative importance of the phytoplankton and (2) evaluating the entire contribution of algal biomass within the context of the Goldsborough & Robinson conceptual model. Our sampling was carried out monthly for a year in sites representative of different conditions with respect to water depth and type of macrophytes. Physicochemical analyses of water were performed following standard methods. Plankton was collected in a five-level profile at deeper stations and in subsurface samples at the shallow one. Samples of sediment obtained with corers were collected for epipelon sampling and segments of plants were cut at different levels, so as to obtain the epiphytes by scraping. Pigment was extracted with aqueous acetone and calculations were made by means of the Lorenzen equation. According to the Goldsborough & Robinson model, a Lake State developed here during the winter (phytoplankton maxima: 150 mg chlorophyll-a per m2). Then, through the subsequent growth of the submerged macrophytes, an Open State was observed, characterized by a maximum epiphyton biomass (at 3,502 mg chlorophyll-a per m2) along with lower levels of phytoplankton biomass. The epiphytic algae on the emergent macrophytes were always present but attained only relatively low biomass values (maximum: 120 mg of chlorophyll-a per m2 in February). The epipelon biomass varied between 50 and 252 mg chlorophyll-a per m2, registering a considerable contribution of settled algae from the water column (phytoplankton). This study contributes to our knowledge of wetland dynamics through its assessment of the rapid changes in the relative contributions of both planktonic and attached algae to the total algal biomass within the context of specific environmental factors.

Ó Springer Science+Business Media B. V. 2008 Abstract Biomass assessments of algae in wetlands usually include only the phytoplankton community without considering the contribution of other algal associations to total algal biomass.This omission prevents an accurate evaluation of the phytoplankton community as an integral part of the total ecosystem.In the present work, the biomass contributions (expressed as chlorophyll-a content per m 2 of lake) of phytoplankton, epiphyton on both submerged and emergent macrophytes, and epipelon were measured in Lacombe Lake, Argentina, for the purpose of (1) establishing the relative importance of the phytoplankton and (2) evaluating the entire contribution of algal biomass within the context of the Goldsborough & Robinson conceptual model.Our sampling was carried out monthly for a year in sites representative of different conditions with respect to water depth and type of macrophytes.Physicochemical analyses of water were performed following standard methods.Plankton was collected in a five-level profile at deeper stations and in subsurface samples at the shallow one.Samples of sediment obtained with corers were collected for epipelon sampling and segments of plants were cut at different levels, so as to obtain the epiphytes by scraping.Pigment was extracted with aqueous acetone and calculations were made by means of the Lorenzen equation.According to the Goldsborough & Robinson model, a Lake State developed here during the winter (phytoplankton maxima: 150 mg chlorophyll-a per m 2 ).Then, through the subsequent growth of the submerged macrophytes, an Open State was observed, characterized by a maximum epiphyton biomass (at 3,502 mg chlorophyll-a per m 2 ) along with lower levels of phytoplankton biomass.The epiphytic algae on the emergent macrophytes were always present but attained only relatively low biomass values (maximum: 120 mg of chlorophyll-a per m 2 in February).The epipelon biomass varied between 50 and 252 mg chlorophyll-a per m 2 , registering a considerable contribution of settled algae from the water column (phytoplankton).This study contributes to our knowledge of wetland dynamics through its assessment of the rapid changes in the relative contributions of both planktonic and attached algae to the total algal biomass within the context of specific environmental factors.

Introduction
Total algal production is usually assessed by taking into account only the phytoplankton community without considering the other algal communities.This omission prevents an evaluation of the contribution of the phytoplankton within the context of the entire ecosystem.In this way, Goldsborough & Robinson (1996) proposed a conceptual model (the G & R Model) that involves different equilibrium states in which the major contributor to the algal biomass is one of the following communities: phytoplankton, epiphyton, metaphyton, and epipelon.This conceptual model has not been extensively tested probably because of the difficulties in measuring accurately the surface area of the submerged and emergent macrophytes to which epiphytic algae are attached.Most studies measure phytoplankton abundance with respect to volume and benthic algae on the basis of substratum surface area, but they do not convert both of these normalizations into an expression of biomass per m 2 of habitat area.
According to Goldsborough & Robinson's (1996) statement: ''the dominant algal assemblage is determined by natural grazing pressure and water column stability and by anthropogenic nutrient loading, water level control, and biomanipulation.These combined, interacting factors determine the duration that a specific state will persist'' In Lacombe Lake, two of the four states described in the G & R model were recorded in a single annual cycle, even though Goldsborough & Robinson (1996) had postulated that each state could persist for decades.Especially in view of the importance that the phytoplankton may acquire in Pampean shallow lakes (Solari et al., 2002(Solari et al., , 2003a, b), b), there is to date only a scant amount of investigation dealing with this algal community within a theoretical context.
From the limnological point of view, the Pampean plain is interesting because of the numerous shallow lakes that are dispersed along the transitional area between the subtropical and Patagonian regions of Argentina.These water bodies constitute a large reservoir for a diversity of fauna and flora (Gabellone et al., 2003).Despite their importance, the lack of information about these wetlands has hindered even a minimal conceptualization about processes within these ecosystems, which understanding in turn is necessary for the development of tools for their management and restoration (Quiro ´s et al., 2002).
Since the Pampean region is characterized by the presence of numerous shallow lakes colonized by submerged and emergent macrophytes, the importance of this research resides in the fact that ours is one of the few investigations including the contribution of all algal communities to the total algal biomass.Presumably, the reason why these kinds of studies are scarce is the great amount of work required for the analysis of the epiphytic communities, taking into account both environmental heterogeneity and different scales of observation.
The G & R model may be used to explain algal biomass dynamics in Pampean shallow lakes mainly because suitable conditions for the development of the different algal communities-such as changes in water level, water-column stability, macrophyte growth, natural grazing pressure, and/or nutrient loading-may occur in those water bodies.
This present paper is aimed at comparing the contribution of phytoplankton biomass at the monitored stations relative to the other algal components within the ecosystem by means of equivalent unitsi.e., chlorophyll-a expressed per m 2 of lake-and evaluating this normalized parameter within the context of the Goldsborough & Robinson conceptual model.

Study area
Lacombe Lake is located in the Pampean region (35850 0 S, 57853 0 W) of Argentina.Lacombe is a seepage shallow lake having an area of 130 ha, a maximum length of 1,750 m, and a maximum width of 1,500 m.The shoreline length is approximately 5.6 km (Fig. 1).During the sampling period, the maximum and minimum depths were 2.5 and 1.60 m, respectively.Emergent and submerged macrophytes (Schoenoplectus californicus (C. A. Meyer) Soja ´k, Potamogeton pectinatus Linnaeus, and Myriophyllum quitense (Kunth) grew, colonizing most of the lake area.This lake was selected owing to its low human disturbance (e.g., extensive cattle breeding).
The climate in this region is temperate without a dry season and with a mean annual rainfall of 1,067 mm (for 1977-2002).March (102 mm), October (102 mm), and November (105 mm) were the historical rainy months during these years.
Dry and wet periods may alternate for decades, influencing the water level of the shallow lakes and thus the ecological dynamics.
The winds at this site blow predominantly from the northeast, south-southwest, and east-southeast directions with an average speed of 11 km h -1 .

Materials and methods
Precipitation measurements were made daily at the Lacombe meteorological station.The samples were taken at 3-or 4-week intervals from July 2001 to June 2002.Four sampling stations (St.1-St.4) were established (Fig. 1): two were located at deep sites among stands of the emergent macrophyte S. californicus (St.1: at the center of the stand, thus minimally exposed to wind action, mean depth of 2.17 m; St. 2: at the periphery of the stand, thus maximally exposed to wind action, mean depth 2.29 m), while two were situated in zones without emergent plants (St.3: at deep sites, mean depth 2.28 m; St. 4: at shallow sites, mean depth 1.15 m).
At each site, water levels were measured with a ruler.Transparency was estimated by means of a Secchi disk.
Water temperature, pH, turbidity, conductivity, and dissolved-oxygen concentration were recorded simultaneously by means of a Horiba U 10 multimeter having a vertical profile (of five levels) at St. 1-St.3, but making only a subsurface measurement at St. 4 owing to its shallowness.Within the vertical profile, water samples for chlorophyll-a and chemical analyses were collected into two 1-l acid-cleaned polyethylene bottles by means of a suction pump during mid-morning and then transported to the laboratory in an isolation box at 5-8°C for subsequent storage in the dark at that temperature until the time of analysis.
The concentration of total suspended solids was determined by the gravimetric method (Method 2540 D, APHA, 1995) and the concentrations of particulate organic matter by the method of weight loss on ignition at 550°C (method 2540 E, APHA, 1995).Total phosphorus (TP) content was determined by the ascorbic acid method after digestion with acidic persulfate (method 4500-P B, APHA, 1995).For phytoplanktonchlorophyll-a determination, a volume of 500 ml was filtered through Whatman GF/C filter and frozen.

Epipelon
Two sediment samples from each sampling location were collected by means of an acrylic corer of diameter 3 cm and height 20 cm with a retention valve and a 2-m length stainless steel tubing.The supernatant water was then carefully removed with a pipette bulb (Robinson, 1983) to avoid the inclusion of planktonic algae.The sediments were next extruded out with a piston in order to obtain the upper 2 cm (Eaton & Moss, 1996), which fraction was then transported to the laboratory in an isolation box at 5-8°C for subsequent storage in the dark at that temperature until the time of analysis.
An aliquot of the sample (25% of the wet weight of each homogenate) was filtered with GF/C filters, with at least two filters being required for each aliquot.The bottle with the aliquot and the filter container were both rinsed with tap water to remove all the sediment adhering to the walls.Before analysis for chlorophyll, the filters were inspected under a binocular microscope and the macrophyte debris removed.The filters were finally stored frozen for later chlorophyll-a analysis.

Epiphyton on emergent macrophytes
The samples of the stems of S. californicus were obtained by cutting sequential 7-cm sections beginning at 1 cm under the water surface and discarding this first 1 cm along with every subsequent second section along the stem.Six to seven sections per stem were sampled according to the depth for each sampling occasion.Every sample was placed in a flask with tap water and then transported to the laboratory in an isolation box at 5-8°C for subsequent storage in the dark at that temperature until the time of analysis.
The stems were scraped with a glass coverslip and the freed material collected in a Petri dish to obtain the epiphyton (Aloi, 1990).The samples were collected on Whatman GF/C filters and the latter frozen for later analysis.The area of each section of stem was determined by the truncate-cone formula and expressed in cm 2 .The amount of chlorophyll-a estimated for each measured section was multiplied by 2 in order to extrapolate this known value to the first half of each adjacent discarded section and the sum of the data for all of the sections-measured and unmeasured-considered to be the total chlorophyll-a contained in the epiphyton on the stem.These chlorophyll-a values were normalized to 1 m 2 of lake area containing bulrush (St. 1 and St. 2) by multiplying this empirical value by the mean density of S. californicus at that location, which figure was obtained by measurement of the stem density of this species in ten random samples from one square meter of each sampling site on a given occasion.

Epiphyton on submerged macrophytes
The apical 10-cm portion of macrophytes was excluded because the colonization time on this section-it being the youngest part of the planthad been brief.A sample (in duplicate) from the next 20 cm was placed in a flask with tap water and transported to the laboratory in an isolation box at 5-8°C for subsequent storage in the dark at that temperature until the time of analysis.
In a Petri dish, all leaves from each sample were removed with dissection forceps and scraped with a glass coverslip.The completeness of removal of the attached algae was monitored by binocular microscopy.The algae and water were filtered through Whatman GF/C filters and the latter frozen for subsequent chlorophyll-a analysis.The dry weight (at 105°C to constant weight) of each scraped, epiphytefree macrophyte was then measured (Sand-Jensen & Sondergaard, 1981).The mean percent coverage of the beds by plants was estimated in ten random 1-m 2 samples on each sampling occasion and site.To estimate the number of plants per unit of area, 30 plants of each species were immersed in tap water in graduated tubes and the area occupied by each plant was measured.The number of plants was then calculated taking into account the mean percent coverage and area occupied for each plant.That the position of the plants estimated by the experimental measurement performed in the laboratory coincided with their actual spatial position in the lake was confirmed by oblique photographs obtained at each sampling site on the corresponding date-with, however, the sole exception of January and February at St. 4, when the plants there showed an incidence of flattening.The values for chlorophyll-a could therefore be underestimated by the calculations for these sampling times at St. 4.
Of forty macrophytes of each species, representing the entire sampling period, the epiphytes were first removed both by manual means with forceps and by ultrasound and then dried at 105°C to constant weight for estimation of the dry weight-length ratio.
Dry weight versus length regression equations were performed for estimation of the dry weight of the macrophyte area occupied by the epiphyton.This calculation was made only over the first 50 cm since observations with the unaided eye and by microscopy of several pieces along all the sampled macrophytes had indicated that the epiphytes colonized mainly this section.Finally, the chlorophyll-a content of epiphytes per macrophyte dry weight was multiplied by the mean dry weight of macrophyte area covered.This latter value was, in turn, calculated as the total macrophyte dry weight multiplied by 1/100 times the mean percent coverage.The results of the chlorophyll-a values per unit of area were obtained for the submerged macrophyte species present in each sampling sector and occasion.

Chlorophyll a
The Whatman GF/C filters with phytoplankton, epipelon, and epiphyton from submerged and emerged macrophytes were incubated in 90% (v/v) aqueous acetone for 48 h in order to perform spectrophotometric measurements (APHA, 1995).The values for chlorophyll concentration thus obtained were corrected for pheopigment concentration by spectrophotometric readings before and after acidification with HCl (Marker et al., 1980).The chlorophyll-a concentration of phytoplankton was calculated according to Lorenzen (1967) and the chlorophyll content of epipelon and epiphyton by a modification of the method of Varela (1981).
Secchi values and phytoplankton chlorophyll-a concentrations were used to obtain light-attenuation coefficients (E), while the euphotic depth (Z eu ) was calculated according to Scheffer (1998, pp. 25, 30).

Statistics
The set of data had been previously transformed and standardized to multivariate analysis.Principal Component Analysis (PCA) between phytoplankton chlorophyll-a concentration per volume and environmental variables was performed by means of a MVSP program (Kovach, 2001).Analysis of Variance (ANOVA) and the least significance difference (LSD) test were carried out to examine variations in the amounts of phytoplankton chlorophyll-a per unit area between different dates and sites.

Results
The sampling period was the rainiest of the previous 32 years (1,336 mm in 2001 and 1,375 mm in 2002), with 1,067 mm being the historical mean.During the sampling period, three major events of heavy rainfalls occurred: August (191 mm), October (223 mm), and March (the largest one: 515 mm).Increments in hydrometric level were markedly related to these events (Fig. 2A).As a result, major changes in water conductivity occurred.During this period, a mean value of 4,140 lS cm -1 was recorded in July, but a sustained decrease in this parameter was observed throughout the sampling period thereafter, reaching a minimum value of 1,244 lS cm -1 at the end.A slight increment in conductivity was observed in the summer months related to a water-level decrease.Water transparency ranged from 0.3 to 1 m, and considerable increases were recorded in coincidence with the October and March precipitation events (Fig. 2B).
The average water temperature was 17.6°C (Table 1), with a maximum peaking in December at the shallower station (30.6°C) and a minimum occurring in winter at St. 1 (9.7°C).The mean dissolved-oxygen concentration was 8.6 mg l -1 (Table 1), with similar values being obtained at different depths along the various vertical profiles during the first 3 months, but with the widest range of variation with depth taking place in summer.An alkaline pH was recorded throughout the entire sampling year.The mean concentrations of the suspended solids were notably high during the winter (mean value for the first 3 months: 67 mg l -1 ), but underwent an abrupt decrease in October (19 mg l -1 ).Total phosphorus concentrations increased markedly during the first two precipitation events (in August and October) but then decreased with the rainfalls in March.The average TP in a water column during the turbid phase was 317 lg l -1 .After a considerable phosphorus input from runoff in October, the mean TP concentration diminished gradually throughout the submerged-macrophyte and epiphyton growth period, from 579 lg l -1 to the minimum of 111 lg l -1 in March (Table 1).
According to the light-attenuation coefficient (E), a marked change in underwater light climate was observed in October, when water transparency increased notably.As a result of higher E values in the first three sampling months, the euphotic zone did not extend to the bottom in the deeper sites (St.1-St.3).From October to December, different conditions of light availability could be observed at the bottom in these sites.Subsequently, light reached the bottom on almost all sampling occasions.In contrast, in the shallow station St. 4, although similar values of E were obtained, the estimated euphotic zone always extended further than the bottom, and consequently the sediments had a good light climate throughout the entire year.
The highest mean concentration of particulate organic matter (POM) in the water (59 mg l -1 ) and the highest mean turbidity (101 NTU) were recorded during the first 3 months (Fig. 3).A sharp decrease in these variables occurred in October (14 mg l -1 POM and 30 NTU, respectively).Coincidentally, P. pectinatus and M. quitense started to grow at stations with less exposure to the wind (St. 1 and St. 4).The maximum submerged-macrophyte coverage at sites with emergent plants was reached in January (St.1: 50%) and in June (St.2: 24%) and at stations without them in February (St. 3 and St. 4: 58 and 100%, respectively), with coverage values at St. 4 being the highest on most occasions.The biomass of submerged macrophytes decreased abruptly in March, but was seen to have recovered by May (Fig. 3).
The density of S. californicus was higher at St. 1 (mean density: 35 axes m -2 ) than at St. 2 (mean density: 10 axes m -2 ) during the sampling period.In both stations the maximum coverage density-T 44 and 17 axes m -2 , respectively-was reached in February.
Mean phytoplankton-chlorophyll concentrations were highest during the first 3 months (range: 43 lg chlorophyll-a l -1 in July to 63 lg chlorophyll-a l -1 in August).Although a notable reduction in pigment average values occurred in October (8 lg chlorophyll-a l -1 ), chlorophyll concentrations peaked in November Fig. 2 (A (56 lg chlorophyll-a l -1 ).From December on, mean values ranged from 27 lg chlorophyll-a l -1 in that month to 9 lg chlorophyll-a l -1 in June.Opposite trends of S. californicus epiphyton chlorophyll were observed, with the lowest recorded mean values occurring in the first 3 months (2-14 mg chlorophyll-a m -2 substrate) and higher ones following during the rest of the period (21-100 mg chlorophyll-a m -2 substrate).The highest pigment concentration was reached in October, but values of about one-half of this maximum were common from November.Similar values and temporal variations were recorded at St. 1 and St. 2, despite the differences in stem densities, submerged-macrophyte coverage, and wind exposure characteristic of these stations.Submerged-macrophyte epiphyte-chlorophyll concentration increased throughout the macrophyte growing season.Nevertheless, in January relatively low mean values were recorded.During the study period pigment values were similar between P. pectinatus and M. quitense, except for February, when considerable differences between the two were observed (Table 2).The amount of phytoplankton chlorophyll per area of lake was higher during the first 5 months (average: 91 mg chlorophyll-a m -2 of lake) with the notable exception of the October measurements, when a considerable decline was observed (average: 17 mg chlorophyll-a m -2 of lake).From December on, this parameter varied between 6 and 120 mg chlorophyll-a m -2 of lake (average: 42 mg chlorophyll-a m -2 of lake) (Fig. 4).
With respect to epipelon chlorophyll, the widest range of values was recorded at the shallow station St. 4 (30-252 mg chlorophyll-a m -2 of lake) (Table 2), where the maximum was reached during January.
Similar fluctuating temporal patterns were found at St. 2 and St. 3, with these locations generally attaining  higher values at the latter station (St.2: 60-150 mg chlorophyll-a m -2 of lake and St. 3: 92-159 mg chlorophyll-a m -2 of lake).The lowest values were observed at St. 1, on almost all sampling occasions, with the maximum here occurring in November (122 mg chlorophyll-a m -2 of lake) and values otherwise varying between 38 and 84 mg chlorophyll-a m -2 of lake for the rest of the period (Fig. 4).The amounts of epiphyton chlorophyll per area of lake on S. californicus were higher at St. 1 (average: 31 mg chlorophyll-a m -2 of lake) than at St. 2 (average: 9 mg chlorophyll-a m -2 of lake), mainly owing to the greater axis densities in the center of this stand.Two peaks were recorded at both sites: in October (61 and 14 mg chlorophyll-a m -2 of lake, respectively) and in February (125 and 27 mg chlorophyll-a m -2 of lake, respectively).During the first 3 months, however, in spite of the large substrate availability, the amounts of epiphyton pigment per area were negligible as a result of the low values of chlorophyll per area of substrate (average: 1 mg chlorophyll-a m -2 lake at both stations; Fig. 4).
Compared to the other algal communities, much higher concentrations of epiphyton chlorophyll were observed on submerged macrophytes in December at all stations (maximum values for St. 1, St. 2, and St. 4) and in February (maximum value for St. 3).Increases in macrophyte coverage and in the amount of chlorophyll per substrate area produced the first peak, while a rise in pigment concentration (per substrate area) on only M. quitense was responsible for the second peak (Fig. 4).
In order to determine if two different phases could be distinguished, a PCA analysis was performed taking into consideration certain physical and chemical variables as well as the phytoplankton-chlorophyll concentrations (per volume of water; Fig. 5).On the first axis, two groups of variables were separated: chlorophyll concentration, conductivity, and TP on the one hand and transparency on the other (42.6% explained variance).On the second axis, two variables highly related to seasonality-pH and temperature-accounted for most of the variance (22.6% explained variance).An ordering of the samples along the first axis resulted in the discrimination of one clear grouping: samples of the July-September period were related to high chlorophyll concentrations, high conductivity, high TP concentrations, and low transparency, while the rest of the samples were ordered according to a decrease in the first three variables mentioned plus an increase in transparency.These samples also distributed along the second axis, with summer samples being the ones that correlated more closely with higher temperature and pH values.In this way, samples from the July to September period were fundamentally different from those of the rest of the year, thus enabling a separation of at least two distinct states.Through the PCA, we also observed that the samples from St. 4 were dissimilar to those from the rest of the sites on various occasions.
An ANOVA analysis was performed testing these differences, but with the amounts of phytoplankton chlorophyll per area (mg m -2 of lake) being used instead.Statistically significant differences were obtained between sampling stations when this variable was considered (one-way ANOVA: P \ 0.05; F = 3.2; significance = 0.032).Moreover, the LSD test showed that St. 4 differed significantly from the rest of the sampling sites (paired comparisons with St. 4: St. 1: 0.014; St. 2: 0.017; St. 3: 0.014; P \ 0.05).
In view of the LSD results, the absolute and percentage values of chlorophyll-a at the shallowest sampling station were compared to the respective data from the deeper stations.During the first 5 months, epipelon-biomass contributions were either similar to, or much higher than, those of the phytoplankton at deeper stations (St.4; Fig. 6).From December to February, a very important contribution of the epiphytic community was recorded.
Importantly, in the former month, the amount of epiphyte chlorophyll per area at St. 4 attained values one order of magnitude higher than those of the other algal communities (Fig. 6D).From March to May, the epipelon contribution was greater than 50% at the deep stations, whereas epiphyton and phytoplankton biomass were variable.In June, the epiphyton community became predominant because of plant recruitment (i.e., increase in substrate availability).At St. 4, the epipelon contribution was greater than at the deeper stations, whereas the epiphyton dominated from May on (Fig. 6A, C).By contrast, at the deep stations, the epipelon biomass remained almost constant throughout the sampling period, while large variations occurred in the planktonic and epiphytic biomasses (Fig. 6A, B).At St. 4, the epipelon biomass increased from December to February.

Discussion
External or internal factors, or both, can disturb a shallow lake, thus either promoting the transition from one state to another or else making the system oscillate around a preestablished state.Periods of flood or drought or human impact (eutrophication, water-level regulation, and emergent-and submerged-plant removal, among others) are the main external factors Fig. 5 PCA analysis between phytoplankton chlorophyll concentration (per liter) and several environmental variables that can impel a change in the state of a Pampean shallow lake.Lacombe Lake has been disturbed only little by human influence, and because it is a seepage lake, climatic conditions are probably the main forces influencing this particular ecosystem.Three rainfall events disturbed the lake's environment during the study period.As a consequence, the different algal communities were affected by the resulting increase in water level and consequent decrease in conductivity as well as by the changes in water-column stability and nutrient concentrations.
Phytoplankton-chlorophyll concentrations (lg l -1 ) coincided with the range of measurements already cited for shallow lakes (Margalef, 1983;Goldsborough & Robinson, 1996;Quiro ´s et al., 2002).Remarkably enough, the values recorded in the first 3 months and in November were all close to the average that characterized ''turbid shallow-water lakes'' according to Quiro ´s et al. (2002), while the values observed during the rest of the study period were quite similar to the average that is typical of ''clear shallow-water lakes'' given by the same authors.
In Lacombe Lake, Solari et al. (2003b), for the same annual cycle, established that the phytoplankton showed a rich flora of coccal green algae and secondarily of filamentous cyanophytes and diatoms.The maximum densities of these algae were recorded during the first 3 months (40,000 ind ml -1 ), with the cyanophytes being the predominant group.In spring, density values diminished, and coccal unicellular chlorophytes and Cyclotella meneghiniana Ku ¨tzing dominated.Later on, in summer, the algal abundance reached the lowest concentrations, with Microcystis aeruginosa (Ku ¨tzing) Ku ¨tzing and Gloeocapsa dermochroa Na ¨geli being the predominant species.In autumn, a high richness of cyanophytes in particular was recorded, but during winter colonial central diatoms and cryptophytes dominated the phytoplankton flora (Solari et al., 2003b).This flora is typical of shallow eutrophic Pampean lakes (Guarrera, 1962;Yacubson, 1965;Ringuelet, 1972;Izaguirre & Vinocur, 1994a, b;Solari et al., 2002;Solari et al. 2003a).
Taking into account the results detailed above and considering the statistical analysis we performed, we were able to recognize the occurrence of two distinct states of the G & R model in Lacombe Lake during the annual cycle: the Lake State (July-September) and the Open State (December-February).

Lake State
The phytoplankton ranged between 40 and 60% of the total biomass (although in the shallow station St. 4 lower percentages-23 to 32%-were reached).The water column was unstable, the nutrient content was relatively high, the transparency was low owing to phytoplankton density, the epiphytic algal biomass was low, and there were only few dispersed aquatic macrophytes.All these properties are characteristic of the Lake State (Goldsborough & Robinson, 1996).This period was considered a phase dominated by the phytoplankton community, even though epipelonbiomass contributions were either similar to, or much higher than, those of the phytoplankton at deeper stations.By contrast, Conde et al. (1999) found that most of the species recorded in the water column of Laguna de Rocha (Uruguay) were nonplanktonic, having been resuspended from the sediments, and thus concluded that the lake was in a Dry State.We support the notion of the presence of the Lake State here since the deeper stations do not have light availability for the sediments and also because the benthic community was mostly of allochtonous origin.Cano et al. (2003) established that the benthic flora was chiefly constituted by various members of the most abundant planktonic taxa (Lyngbya limnetica Lemmerman, Aphanocapsa delicatissima W. et G. S. West, Spirulina laxissima G. S. West, Oocystella parva (W.et G. S. West) Hinda ´k and Tetraedron minimum (A.Braun) Hansgirg).

Open State
The epiphyton biomass dominated, contributing 53-94% of the total-area biomass.In addition, the most important characteristics of an Open State were: variable water transparency, unstable water column conditions, lower nutrient and phytoplankton-chlorophyll concentrations, and development of submerged macrophytes.The occurrence of this stage was evident in all of the sampling stations, with the highest percentages of epiphyton biomass being observed in the shallowest station.Maximum contribution of the epiphyte community was recorded in December because of the combined effect of extremely high submerged-macrophyte coverage and elevated concentrations of epiphyte chlorophyll-a per area of substrate.

Transitional periods
Two transitional periods-following heavy rainfalls, which probably disturbed the system-could be distinguished.Although the October precipitations, preceding the first of these watershed periods, were not the highest rainfalls, their influence was of primary importance because they brought about the onset of a new state.Changes in the light conditions and the nutrient concentrations favored the development of extended stands of submerged macrophytes along with the alternative growth of epiphyton in October and of phytoplankton in November.No definite state could be assigned to the October-November period, it thus being regarded as a transitional phase between the clearly established Lake and Open states of the surrounding months.In this transitional phase, new stabilizing mechanisms characterizing the Open State (growth of submerged macrophytes) were set into motion.These plants constituted an extensive substrate area for epiphyton development, with those algae being effective competitors with the phytoplankton for both nutrient and light resources (Sand-Jensen & Borum, 1991, Strand & Weisner, 2001).Submerged plants also provided a safe refuge for the zooplankton, mainly for cladocerans (Claps et al., 2005).A large-body-size phytoplankton (colonies of Microcystis aeruginosa (Ku ¨tzing) Ku ¨tzing) was detected in November by Solari et al. (2003b), and its predominance could therefore be a consequence of grazing pressure over the smaller algal species.
Although the second environmental disturbance, the rainfall event in March, was much higher, the ecosystem was not driven into a new state at that time, but rather oscillated around an Open State condition.The water level increased, affecting at first the abundance of submerged macrophytes, but then a fast recovery of the plants was observed.Despite the initial decline in plants, water transparency remained high.In the deeper stations, epiphyte biomass varied according to the changes in plant abundance, becoming dominant only in June, at the end of the sampling period.Although chlorophyll concentration per liter decreased, the phytoplankton attained a high percentage contribution (39%) owing to the increase in the depth of the water column.Moreover, at these sites, epipelon biomass contributions were the highest (57%), but as was explained above, the allochtonous origin of the sediment flora prevented us from considering this period a bona fide Dry State.The situation at St. 4 was quite different, however.A major biomass contribution was made by the epipelon (61-92%), although the amounts of epipelon chlorophyll-a per unit area were similar to those found in February during the Open State.The contribution of allochtonous species was furthermore lower than had been recorded on other occasions, while the development of the autochthonous epipelic flora could be observed (Cano et al., 2003).The shallowness of this site and the improvement in light conditions caused by the decline in macrophytes may have been the principal factors that gave rise to epipelon growth.The recovery of submerged plants began in May, and thus the epiphyton soon became predominant up until the end of the study.
Because of the characteristics that this shallow lake acquired at the deeper stations during this final period, only in June could we consider it to have been in an Open State.The other 3 months may be regarded as a transitional stage generated by large disturbance, when the contributions of epipelon and phytoplankton became the most important.At St. 4 this transitional period occurred during only 2 months; otherwise this station exhibited definite characteristics of a Dry State.Nevertheless, the submerged macrophytes showed a fast recovery afterward, in this way reestablishing Open State conditions early in May.
The Sheltered State (proposed in the G & R Model) was not detected, owing to the constant wind action over the entire water body, which disturbance favored instability of the water column and prevented the development of the metaphyton characteristic of this state.Rather, that perturbation of the water column characterized the Open and Lake States according to Goldsborough & Robinson (1996).
The Lake and Open State conditions were distinguished at all sampling stations and thus could be extrapolated to the entire lake system.By contrast, the Dry State conditions were observed only at the shallowest station, with it representing a mere 12% of the lake area.
The state changes could occur more rapidly (i.e., months) under a moderate climate in contrast to the long-term changes proposed in G & R model, based on studies of wetlands from cold climate.Lacombe Lake is situated at a warm-temperate region, with a clear seasonality, and also exhibits natural changes in hydrometric level and water conductivity over successive annual seasons and during different years as a result of variability in rainfall.Seasonal variation in precipitation and temperature seems to be important in the context of changes in macrophyte and algal growth in the study lake.The year of this sampling period was rainy, but within the usual seasonal pattern.Moreover, the absence of a water source (rivers and streams) to introduce nutrients and an algal inoculum insures that the changes in algal biomass observed were not related to external sources.The G & R Model can be tested here, knowing that all algal growth cycles and the interactions between algal assemblages were completed within the lake.Because of these interacting key conditions, we were able to conclude that rapid changes in the biomass contributions of different algal communities-mainly nonplanktonic algaedetermine the occurrence of different states within a single annual cycle of this wetland.

Fig. 1
Fig. 1 Location of the sampling stations in Lacombe Lake

Fig. 3
Fig. 3 Mean water particulate organic matter (POM) and mean turbidity considering the four sampling stations (lines), and submerged macrophyte cover (bars) at the four sampling stations recorded in the annual cycle

Fig. 6
Fig. 6 Chlorophyll content contribution of the algal communities to the total algal biomass.(A) Mean percentage at the deep stations (St. 1, St. 2, and St. 3), (B) Mean chlorophyll

Table 2
Mean, standard deviation (SD), range and number of observations of phytoplankton, epipelon and epiphyton chlorophyll a concentrations (expressed by unit of substrate and by