Microb Ecol (2002) 43:67±81 DOI: 10.1007/s00248-001-1015-7 Ó 2002 Springer-Verlag New York Inc.

Food Web Structure in the Recently Flooded Sep Reservoir as Inferred From Phytoplankton Population Dynamics and Living Microbial Biomass R.D. TadonleÂkeÂ, L.B. Jugnia, T. Sime-Ngando, J. Devaux, J.C. Romagoux Laboratoire de Biologie des Protistes, UMR CNRS 6023, Universite Blaise-Pascal (Clermont-Ferrand II), 63177 AubieÁre Cedex, France Received: 16 January 2001; Accepted: 28 June 2001; Online Publication: 23 January 2002

A

B S T R A C T

Phytoplankton dynamics, bacterial standing stocks and living microbial biomass (derived from ATP measurements, 0.7±200 lm size class) were examined in 1996 in the newly ¯ooded (1995) Sep Reservoir (`Massif Central,' France), for evidence of the importance of the microbial food web relative to the traditional food chain. Phosphate concentrations were low, N:P ratios were high, and phosphate losses converted into carbon accounted for 150 (mean = 245 ‹ 124.7, range 23.5±578) for dissolved inorganic forms (Fig. 3). In the reservoir, nutrient concentrations and N:P ratios, as well as N-NO2 + NO3 and P-PO4 relative contributions

to the total nutrient pools of interest, were similar to those found in the streams. Exceptions for these observations were the N-NH3 concentration which was higher in the reservoir than at the entry of the reservoir, and the TP load for which the spatial pattern was inverse relative to that of N-NH3. For nitrogen sources and silicate, mean concentrations generally increased in the deepest waters, whereas for phosphorus sources, no marked depth-to-depth differences were observed. The vertical differences were statistically signi®cant for silicate only (Table 1). Temporal changes in nutrient concentrations at the different depths sampled are shown in Fig. 4. For all these depths, N-NH3 concentrations decreased markedly between March and April and remained relatively constant thereafter. N(NO2+NO3), TN, and SiD amounts generally decreased during the study. The temporal pattern for P-PO4 concentrations was similar to that of N-NH3, with relatively constant concentrations in the water column (at about 5 lg L)1) after April, except in May and early June in the epilimnion when P-PO4 was undetectable. TP concentrations varied little in the water column, although occasional peaks were noted in June (epilimnion) and in September (hypolimnion).

Phytoplankton Species Composition and Succession Fig. 3. Temporal variations in the N:P ratios (total and inorganic, by atom) at the entry (i.e., in the incoming streams Sep and Riaux) of the Sep Reservoir in 1996.

A total of 140 phytoplankton species belonging to 3 taxonomic groups, namely chlorophytes, pyrrhophytes (i.e.,

72

R.D. TadonleÂke et al.

Fig. 4. Seasonal variations in the concentrations of nutrients at depths representative of epilimnion (0 and 1 m), metalimnion (7 m), and hypolimnion (15 m) in the Sep Reservoir in 1996. See Table 1 for other sampled depths.

cryptophytes and dino¯agellates), and chrysophytes, were identi®ed, of which 69 accounted, at least once during the study, for 5% of the phytoplankton total biomass (R.D. TadonleÂkeÂ, PhD thesis, Universite Blaise Pascal, ClermontFerrand, France, 1999). Among these 69 species, the temporal ¯uctuations in the relative biomass of those which occurred at least twice during the study are presented in Fig. 5. The chlorophyte species Eudorina sp. (May 22),

Pandorina smithii (May 22), and Stephanoporos sp. (July 3) occurred only once during the study, but each of the three species accounted, at the time of its occurrence, for more than 15% of the phytoplankton total biomass. For this reason, the three species are included as separated items in Fig. 5. Until April (i.e., the winter±early spring period), diatoms (Rhizosolenia longiseta and Asterionella formosa) and cryptophytes (Cryptomonas ovata and

Fig. 5. Seasonal variations in the contribution of the main phytoplankton species to total phytoplankton biomass in the euphotic zone of the Sep Reservoir in 1996.

Food Web Structure in New Reserviors

73

Chroomonas acuta) were the major contributors to the phytoplankton total biomass. In May and June (late spring), colonial species, the chlorophytes Eudorina sp., Pandorina smithii, Volvox aureus, and the chrysophyte Dinobryon divergens developed in the euphotic zone of the reservoir. D. divergens was generally the most quantitatively important species and persisted throughout the late spring period, although P-PO4 concentrations were undetectable. From July to September (summer), diatoms and cryptophytes became dominant. They mainly comprised small size (5 ´ 106 cell L)1) recorded in the epilimnion on May 22 (i.e., with the strong development of Eudorina sp. and P. smithii) and in June (with the development of V. aureus and D. divergens). In winter±early spring (i.e., February±early Table 2.

Simple statistics for biological variables under study

Depths (m) andSeasons 0 1 4 7 10 15 20 Winter±spring

Summer

May) and in summer (July±September), phytoplankton densities in the sampled depths were 10 lg L)1) were recorded in the epi- and metalimnion, with peak values in the epilimnion in late July, late August, and mid-September (Fig. 6B). The increase in Chl a from late spring contrasts with the decrease in TN, N-(NO2+NO3), and SiD concentrations. Phytoplankton carbon biomass (PhytoC) exhibited exceptional peaks of 1316.4 and 1088.4 lgC L)1 on May 22 at 1 and 4 m, respectively, i.e., with the emphatic development of Eudorina sp. and P. smithii (Figs. 5, 6C; Table 2). When excluding these peaks, PhytoC averaged 99.2 ‹ 91.2 lgC L)1 for discrete depths, and 748.9 ‹ 421.1 mgC m)2

PhytoA (´ 106 cell L)1) 0.24±14.6 4.1 (3.2) 0.92±11.4 4 (2.9) 0.5±6.8 3.1 (1.6) 0.35±3.8 1.8 (1.0) 0.15±2.15 1 (0.6) 0.12±1.95 0.5 (0.5) 0.1±1.35 0.5 (0.3) 021±14.6 2.49 (1.98) 5.6±46.4a 232 (15.3)a 0.1±6.8 1.93 (1.43) 4.1±26a 15.6 (5.95)a

Chi a (lg L)1)

PhytoC (lgC L)1)

0.91±22.5 10.08 (6.06) 0.91±21.3 9.81 (5.77) 1.19±19.86 8.94 (5.78) 0.76±14.9 6.21 (4.25) 1.14±12.8 4.17 (3.67) 0.47±9.47 2.27 (1.99) 0.51±7.81 1.86 (1.67) 0.5±11.77 3.49 (3) 92±46.3b 27.8 (13.7)b 0.47±22.5 8.37 (6) 31.3±80.5b 59.6 (14.14)b

22.8±428.4 143.6 (108.6) 16.2±1316.4 196.42 (275.2) 26.4±1088.4 188.1 (231.6) 132±315.6 108.4 (79.4) 7.2±212.4 72.8 (55.7) 3.6±252 36.57 (57.2) 2.4±124.08 26.76 (32.99) 5.64±1316 135.3 (223.7) 355±60972c 1492.1 (1911.3)c 2.4±428.4 101.69 (97) 128.7±1331.9c 678.8 (368.94)c

BactA (´ 106 cell mL)1)

BactC (lgC L)1)

0.99±8.3 4.11 (2.36)

11.62±122.97 59.62 (37.05)

1.06±5.2 2.44 (1.1)

16.78±111.29 44.84 (26.73)

0.9±2.89 1.54 (0.49)

17.24±87.19 26.59 (59.57)

0.9±327 1.79 (0.7)

11.29±43.75 24.34 (8.1)

1.19±8.3 3.48 (2.1)

16.39±122.97 59.42 (32.4)

LMC (lgC L)1) 246.25±1682.5 542.6 (344.1) 278.3±2300 577.8 (484.3) 266.7±2075 513.9 (497.9) 125±1150 379.6 (244) 65±687.5 229.2 (171.4) 62.5±800 240.17 (193.05) 50±385 193.07 (111.63) 86±2300 438.96 (451.8) 943.3±13495.6d 4072 (4262.1)d 62.5±1150 408 (248.6) 1220.3±2525d 1863.9 (450)d

For each depth and season, the range of values is given on the ®rst line and the mean value (‹sd) on the second line. For each season, two additional lines give ranges and mean (‹sd) euphotic zone integrated values (a ´ 106 cells m)2, b mgChI m)2, c,d mgC m)2).  Results of the Kruskal±Wallis analysis of variance and comparison test (signi®cance level = 5%) indicated signi®cant differences between the sampled zones, i.e., epi (0, 1, or 4 m), meta (7, 10 m) and hypolimnion (15, 20 m).

R.D. TadonleÂke et al.

74

whereas P-PO4 concentrations were undetectable. This period was characterized by low Chl a concentrations in the epilimnion and the euphotic zone of the reservoir, a seasonal pattern that, indeed, contrasts with that of PhytoC (Fig. 6B±D). Heterotrophic Bacteria Bacterial abundance (BactA) and carbon biomass (BactC) in the sampled depths averaged (‹ SD) 2.8 ‹ 1.8 ´ 106 cells mL)1 and 44.6 ‹ 30.5 lgC L)1. The mean values for both variables decreased signi®cantly with depth, apparently due to vertical differences associated with the thermal strati®cation period (Fig. 7A, B, Table 2). Changes in BactA and BactC were generally similar. In the winterspring period, both variables were at their lowest levels (