Phycology, sedimentology, and paleolimnology near Cape Maclear, Lake Malawi, Africa  

Southern Lake Malawi was sampled thoroughly to determine algal distributions, the effects of sedimentary process, and paleolimnology. Rocks were covered with mats of the chlorophyte Cladophora or the cyanophyte Calothrix, accompanied by the diatoms Rhopalodia, Cymbella, and Navicula. Common diatoms among sand grains were Rhopalodia, Navicula, Epithemia, Fragilaria brevistriata, and Surirella.

The planktonic biomass was dominated during the stratified (calm) season by Peridinium, Anabaena, and Nitzschia cf. spiculum. In the mixing (windy) season, diatoms dominate the biomass, including approximately equal cell abundances of Stephanodiscus, Nitzschia cf. spiculum, and Melosira. The chlorophyte Oedogonium was also common.

Sediment trap collections revealed strong misrepresentation among diatoms. Melosira accounted for 32% of planktonic diatom cells, but 57% of the diatoms in the deeper (29m) sediment trap (Kruskal-Wallis test, p < 0.005). Nitzschia cf. spiculum was dramatically under-represented in traps: 53% of planktonic diatom cells but only 14% in the deeper trap (p << 0.005). These data demonstrate that sedimenting diatom assemblages are sometimes a distorted representation of the living community.

The surface sediment assemblage was also different from the plankton. Melosira rose from 32% in the plankton to 53% of the deposited assemblage (p < 0.005). The abundance of Nitzschia cf. spiculum plummeted from 53% of the plankton to 0.8% in the surface mud (p < 0.005).

Grab samples were used to develop transfer functions for paleodepth estimation. The microfossil assemblage was more strongly correlated with depth rather than distance from shore. Seven depth functions resulted in r² values exceeding 0.59 and 95% confidence limits of 5 to 27 m; the best was total percent planktonic diatoms (r² = 0.73, ± 7-16m).

These paleodepth transfer functions were applied to Core MK-1, and the results match the historical record rather poorly. Because of its strong dependence on the southeasterly trade winds, Melosira suggests relatively consistent wind strength since about 1700 A.D. Benthic diatoms show little change as well, so it seems unlikely that lake level has varied substantially (i.e. <20 m) from modern since 1700.

In Core 6P, Melosira was relatively uncommon in sediments representing approximately the third and fifth millenia BP. Local upwelling must have been quite reduced. Abundant Stephanodiscus there suggests that southerly trade winds were weaker, not absent. The benthic diatom Rhopalodia may indicate a slight lowering of lake level in the last two to three millenia. Therefore, the predominant signal recorded in the sediments is one of varying wind strength. Lake level apparently has not been more than 40m lower than modern in the last six millenia.

Large-lake paleolimnologists must not assume fair representation of diatoms in sediments. Major misrepresentations cause Nitzschia to be virtually absent from surface sediments, although it is the most common genus among the planktonic diatoms.