Spatial patterns and trophic modelling within the Lake Malawi Demersal fish community : conservation and fisheries applications

Darwall, William Robert Thomas

Biological sciences
September 2003

Thesis or dissertation

© 2003 William Robert Thomas Darwall. All rights reserved. No part of this publication may be reproduced without the written permission of the copyright holder.

Lake Malawi has a complex, highly speciose, fish community supporting a diverse fishery of great importance as a food source for subsistence communities and to local and national economies. The fish community is also of international importance to scientists interested in the study of evolutionary mechanisms. Little information is available on the structure and ecology of the demersal fish community to help managers understand how best to both sustain fisheries productivity and maintain species diversity. Community level impacts of the demersal trawl fishery in the southern part of the lake are documented but mechanisms for the impacts are unknown. A greater understanding of species distributions and the trophic structure and function of the demersal fish community are required.

The principal aim of this thesis was to improve the information base on demersal fish distributions for use in the selection and design o f aquatic reserves, and to enable a more predictive approach to management of the fish community through ecosystem modelling. Specific objectives were: describe spatial patterns of species distributions throughout the lake; determine the trophic structure of the fish community through dietary studies; model community trophic structure using the ECOPATH software, and; evaluate the ECOSIM simulation software for predicting impacts of ecosystem disturbance such as from demersal trawling and the introduction of non-native species.

Species catch compositions from over four hundred demersal trawls conducted throughout Lake Malawi were analysed for spatial distribution patterns. Species assemblages were differentiated along gradients of water depth and geographic location. Depth differentiation was characterised by a combination of species replacements, differences in the abundance of shared species, and a reduction in species richness with increasing depth. Only the catfish species were shared by species assemblages from all depths. Assemblage similarity along the geographic gradient was negatively correlated with the distance between sample areas but the proportion of species shared between sample areas showed no clear spatial pattern. A number of hypotheses are proposed as possible explanations for the evolution and persistence of the observed pattern of species assemblages. In particular, the lake-wide consistency in species richness at depth is discussed in the context of species-energy limitations within the system. Finally, the potential application of information on species spatial patterns to the selection of ‘no-take’ conservation areas and aquatic reserves is discussed.

Diet analysis, supported by analysis of stable isotope ratios, identified ten main trophic guilds among the demersal fish community ranging across three to four trophic levels. An additional guild was formed to incorporate those shallow water species not often encountered in the demersal trawls that were reported in the literature as feeding on epilithic algae and macrophytes. The mean trophic level o f the demersal fish community for all depths and sample areas, pooled and weighted by the respective areas of lakebed coverage, was estimated at 3.18 with the bulk of biomass (58%) concentrated at trophic level 3.0. Mean trophic levels of fish assemblages were significantly influenced by water depth and sample location. The main influence of depth was realised in the shallow water where mean trophic values in the 0 to 20 m depth band were significantly lower than at all other depths.

For many taxa within the demersal fish community there appeared to be a high proportion of trophic-equivalence. It was therefore thought that the community included a number of functional analogues that might be maintained by spatial variability, or perhaps by unobserved behavioural mechanisms operating across species complements. While subtle mechanisms may be at play to support the unusually high species diversity of the lake the results suggested that elimination of a species might result in its subsequent replacement by a functional analogue. Caution should be exercised, however, in making this interpretation based on the rather broad dietary characteristics considered here. Moreover, such functional analogues could also be highly important for the maintenance of ecosystem function and stability through provision of a buffer against environmental change. Loss of these species could thus be highly damaging to the long-term resilience and ecological integrity of the community and, consequently, to the viability of the fishery.

Energy flows between trophic levels were estimated through food consumption studies for a selection o f representative species. A trophic model o f the demersal fish community and a graphical representation o f the food web and its energy flows were constructed using the ECOPATH approach and software. The area of the lake modelled was restricted to the southern and western sectors o f the lake to a water depth of approximately 140m.

The proposed role of the lakefly (Chaoborus edulis) in the ecosystem was somewhat clarified. Previous study showed that C. edulis was a more important food source to the pelagic fish community than originally suggested, and went on to propose that demersal fish might also rely on consumption of C. edulis larvae. Integration of the demersal and pelagic systems to form a model of the complete ecosystem showed that approximately 57 % of C. edulis production was directly consumed by pelagic fish and a further 2.4 % by demersal fish. The remaining 40 % of production either flowed to detritus, where a certain proportion was recycled through detritivores, or left the system through dispersal as flying adults, some of which returned to the lake to be consumed by surface-feeding catfish.

In the demersal zone the role of C. edulis was less significant than in the pelagic. System modelling suggested that although C. edulis provided a direct link between the demersal fish community and pelagic productivity the main pathways for energy flow in the demersal domain were through the consumption of detritus and, to a lesser extent, consumption of copepods by demersal fish apparently migrating into the pelagic for feeding. Integration of the demersal and pelagic components of the Lake Malawi ecosystem indicated that, for the area of lake modelled, the system was more efficient than previously supposed as the demersal fish community was able to utilise much of the pelagic production. In summary, the demersal system appeared to rely most heavily on biomass imported from the pelagic through consumption of detritus and of copepods consumed by fish migrating vertically to feed in the pelagic.

Sensitivity analyses were run for a number of ECOSIM simulations based on disturbance scenarios of relevance to the Lake Malawi demersal fish community. Simulations included increased fishing yields, initiation of a deep-water trawl fishery, and the introduction of an alien pelagic zooplanktivore. In all cases the quantitative nature of ECOSIM predictions was sensitive to input values for the ‘vulnerability’ parameter while the qualitative nature of predictions was unaffected. As available data were found insufficient to obtain robust estimates for vulnerability it was concluded that ECOSIM simulations should not be employed to predict the quantitative nature o f group biomass changes but may be used in a qualitative way to predict the direction of change.

In all impact scenarios modelled ECOSIM simulations predicted a change in species composition and community trophic structure. It was suggested that the predicted changes in trophic structure might lead to unpredictable outcomes such as trophic cascades and abrupt phase shifts in community structure with associated reduction or stagnation of catches.

In conclusion, this study has demonstrated the complex nature of trophic interactions within the Lake Malawi fish community. Modelling simulations have demonstrated the potential for human intervention to disrupt community structure with associated consequences of reduced productivity and stability of the fishery. A number of gaps in knowledge have been identified which, if filled, would greatly increase the robustness of the model and increase its utility as a predictive tool for managing the fishery.

Department of Biological Sciences, The University of Hull
Turner, George F.
Sponsor (Organisation)
European Community Fourth Framework Programme
Ethos identifier
Qualification level
Qualification name
27 MB
QR Code