Synthesis of simple predictive models for river fish yields in tropical lakes

Background

For many low-income groups in developing countries, fish is the main source of animal protein in the diet. Much of the fish consumed comes from small inland water bodies, which are individually not important enough to merit fisheries surveys, especially as the resources of Government Fisheries Departments are usually very limited. Without fisheries survey data, managers cannot recommend measures to ensure that fish populations are managed sustainably, and without resource inventories, they cannot prioritise allocation of resources for sectoral development. Estimates of potential sustainable fish yields are not only valuable for sound management of existing fisheries, they are also useful when planning and evaluating the likely benefits and impacts of constructing new reservoirs for irrigation, electricity generation or domestic water supply.

Approach

This project aimed to provide a simple method of estimating the fishery potential of lakes and reservoirs by establishing predictive relationships between fish yields and simple environmental, climatic and demographic parameters. This was accomplished using published information from the wide range of tropical and sub-tropical water bodies that have been studied or monitored to some extent.

A primary database was constructed containing annotated statistics on morphology, hydrology, chemistry, biology, fisheries and catchment demography for tropical and sub-tropical lakes, reservoirs, swamps and coastal lagoons. The database provides scope for empirical analysis of a variety of scientific hypotheses concerning the role of various factors affecting the productivity of aquatic ecosystems.

A summary table within the database was used to explore relationships between fish catch and a range of parameters in inland water bodies (excluding rivers, see R5030), to produce simple predictive models.

Findings

A flexible database, relating physical, demographic, climatic and fisheries characteristics of tropical lakes and a number of models for predicting lake yields from simple physical or biological data were produced.

For example, simple, broadly applicable relationships between fishery yield, fishing effort (as number of fishers) and water body area and volume explain much of the variability in total catches on a global and continental scale. While relationships between water body area and yield, and between the number of fishers and catch, do not differ significantly between continents but these relationships have a high variance for Asian and Latin American lakes. More detail on additional models determined can be found within the Final Technical Report.

The models contributed to the FAO Atlas of World Fisheries.