Wat. Sci. Tech. Vol. 32. No.4. pp. 139-142. I99S. Copyrlgbl C I99S IAWQ Prioled in o...at Bnlam. All ngbl& reserved. 0273-1223-"1S S9"50 + 0-00
HORIZONTAL DISTRIBUTION PATIERNS DURING A CYANOBACTERIAL BLOOM K. Pettersson, L. Forsell and A. T. Hasselrot Uppsala University, Institute of Limnology, Erken Laboratory, Norr Malma 4200, S-761 73 Norrtlllje. Sweden
ABSTRACT A hierarchical sampling was perfonned in order to give a picture of the borizonlal dislribution of cyanobacterial colonies. It showed bow the unportanee of different scales in distance cbanged during the summer bloom of Gloeolrichia echiflulala in the moderately euuopbic and suatified Lake Erken in southeastem Sweden. G. echiflulala colonies occurred in patcbes at a distanCe of -1000 m and the variance at smaller scales was low. For colony content of phospborus. cbloropbyll a and size. however. the largest vanances were found at the 20 m (and error) scale If the vanances between dates are not taken mto accounL It was concluded that in order to estimate the lake population of cyanobacterial colonies properly these variances must be conSidered when deciding the sampling program.
KEYWORDS Cyanobacteria; horizontal distribution: hierarchical sampling; phosphorus. INTRODUCTION Gloeotrichia echinulata is a large. filamentous and colonial cyanobacterium. It is one of the dominating species in Lake Erken during late summer. Earlier studies have indicated that the inoculum is unusually large and that the colonies epilimnetic uptake of phosphorus is almost nonexistent (Barbiero. 1993; Forsell, 1993; Istvl1novics et a/.• 1993; Pettersson et al.• 1993: Istvanovics et a/., 1994: Forsell and Pettersson. in press).
While looking at the migration and nutrient transport of the cyanobacterium Gloeotrichia echinu/ata we were faced with some questions that initiated the present study. Was there a significant transport of colonies (and nutrients) from the littoral to the pelagic areas and backwards? The facts that the migration of G. echinulata is greater from depths above -4 m (Barbiero. 1(93). and that the number of akinetes in Lake Erken sediments increa.~es with decreasing depth. seemed to indicate this. [f so, would this leave any trace in the horizontal distribution of the epilimnetic colonies? Furthermore. the sampling programme made it possible to accurately determine the size of the population. Therefore. we followed the horizontal distribution of G. echinulata colonies during the bloom period in 1994. The use of the hierarchical sampling programme also made it possible to estimate the correctness of an ordinary sampling where only a few sampling point~ were used. 139
K. PE1TERSSON ~I a1.
Our aim was to determine the horizontal variance, at what scales any variation lay, if the importance of the different scales varied during the bloom and the size of the population. MATERIAL AND METHODS Lake Erken is a moderately eutrophic lake in Southeastern Sweden. It has an area of 24 km 2, a maximum depth of 21 m, a mean depth of 9 m and stratifies during summer. Water samples were taken each week from July 13 to August 23 in 1994. A total of 81 samples were taken randomly on each sampling occasion at four scale levels according to a hierarchical design (Table 1). Each sampling area at one level included three areas at the level below. The whole epilimnetic water column was sampled with a tube sampler and the number of colonies calculated per m 2 of surface area. From August 3, determinations were also made of size, phosphorus and chlorophyll a content on respectively 5, 25 and 100 colonies from each sample. All variables in this study were log+ I transformed prior to analysis. This was done in order to stabilize variance and to avoid correlations between mean and variance in the different hierarchic levels. A random nested ANOVA model (using SAS Institute Inc. 1990) was used to evaluate each levels contribution (variance components) to the total variance. This was done both for the whole period and for the separate sampling days. Table I. The total variance at the different levels Level
Colonies % p
Phosphorus % P
Chlorophyll a % P
20 m (error)
RESULTS AND DISCUSSION For the whole period the largest variances for all the parameters were naturally found between the different days. The variance at the category level was often small and non-significant (> 0.5), Table I. The variance at the different levels is given as % of the total variance components for the whole period (from July 13 to August 23 for colony numbers and from August 3 to 23 for the other parameters), with p-values. The 20m level also includes the sampling error. For the separate sampling days, the largest differences in variance for colony numbers occurred at the 1000 m and category levels (Fig. la). The variances in the latter were often non-significant but indicate a possible difference between the categories. The difference between samples at the 20 m (and error) level was small for colony numbers. However, at this level (20 m) the largest variances occurred for the other parameters (Fig. Ib and c). This was probably mostly due to errors in analysis and the number of colonies used for the different analyses.
This hierarchical method gives a good pic~ of the changes in the horizontal population structure of G. echinulata during the bloom. As these changes probably are lake specific. the method might be wen suited for comparisons over time between different lakes and also for other parameters.
The results are preliminary, and further study of the material will hopefully give some clues to the connections between the different variances over time. 100 ~
20 0 13,7
CQ 'C CQ
[ffill % lO00m •
o % 20 m (error)
40 20 0 3,8
Date FiJlft I. The CODlribulioa 10 v.-iaace from lbe differenl levels over lime. The variallce at lbe dlfferenllevels over lime, . . . of lbe lOlaI VIriaIK:e compoIICIIlS for colony numbers (a). for pbospborus COlIletll • coIooy ·1 (b) and fOf c:bloropbyU a contelll • colony -I (c).
K. PETrERSSON et al.
CONCLUSION We found that a significant horizontal variance exists for G. echinulata in Lake Erken. The greatest variance in colony numbers occurred at distances of -1000 m. While the variance in number was low at the 20 m (and error) scale, it was here that the largest variances were found for colony size and content of phosphorus and chlorophyll a. The importance of the different scales did however vary during the bloom. ACKNOWLEDGEMENT The sampling program and analyses were performed by Katarina Bratt. Kristina JOnsson, Andrea GaInander. Craig Williams and Sonja Ghalib. The economical support was a scientific grant from the Swedish National Protection Agency. REFERENCES Barbiero. R. P. (1993). A contribution 10 the life hisoory of the planktonic cyanophyte Gloeotrichia echinulata. Archive fUr Hydrobiologie. 117 87-100. Forsell, L. (1993). Some aspects of the migration and distribution of the cyanobacteria Gloeotrichia echinulala. 21st Nordic Sediment Symposium, K. Pettersson and L. ForseU (cds.). Institute of Limnology. UppsaJa university reports LIU 1993 B: 7. pp. 105-118. Forsell. L. and Pettersson. K. (in press). On the seasonal migration and its influence on the pelagic population of the cyanobacterium Gloeotrichia echinulata in Lake Erken. Sweden. Marine and Freshwater Research. IstvAnovics. V.. Padisak. J.• Pettersson, K. and Pierson. D. (1994). Growth and phosphorus uptake of summer phytoplankton in Lake Erken (Sweden). J. Plankton Res., 16, 1167-1196. IstvAnovics. V•• Pettersson. K•• Rodrigo. M.• Pierson. P.. Padisak. J. and Colom, C. 1993 Gloeolrichia echinulala. a colonial cyanobacterium with a unique phosphorus uptake and life strategy. J. Plankton Res.• 15, 531-552. Pettersson. K., Herlitz, E. and Istv6novics, V. 1993. The role of Gloeotrichia echinulma in the transfer of phosphorus from sediments 10 water in Lake Erken. Hydrobiologia. 253,123-129.