7. GROWTH AND PRODUCTION CHARACTERISTICS OF CHAMBO WITH A DISCUSSION ON MESH SELECTIVITY OF CHAMBO EXPLOITING GEARS

7.1 ESTIMATION OF GROWTH PARAMETERS

Three species of chambo, Oreochromis lidole, O. squamipinnis and O. karongae, occur in Lake Malawi. The latter two constitute the bulk of the catch in Lake Malombe. The length-frequency data collected from all types of chambo exploiting gears in the south-east arm of Lake Malawi and Lake Malombe between August 1990 and September 1991 form the main data base of this study. See Seisay et al., 1992b for details.

Monthly total catches from the artisanal fisheries, enumerated by gear type, are derived from the Malawi Traditional Fisheries (MTF) catch-effort data collection programme.

Figure 7.1 Length converted catch curves of the south-east arm of Lake Malawi chambo species: O. lidole (A), O. squamipinnis (B) and O. karongae (C). For Lake Malombe O. squamipinnis and O. karongae are combined (D).

Each length-frequency data set was resolved into normally distributed cohort components using the Bhattacharya method (LFSA package, Sparre, 1987).

The Gulland and Holt plot and the Von Bertalanffy non-linear regression method was used to estimate growth parameters from the dissected mean lengths. The Gulland and Holt plot gave K and L_∞. The Von Bertalanffy plot provided another set of k, L_∞ and t₀ estimates. Age-determination, using opercular bones also gave growth parameter estimates (Banda, 1992).

The estimated growth parameters are given in Table 7.1. The Gulland and Holt plot overestimates K compared to the Von Bertalanffy plot and estimates of ageing from opercular bone measurements. The growth estimates from the Von Bertalanffy plot compared well with those from ageing. From their growth performance indices (Phi Prime), the different species of chambo showed similar growth patterns. However, O. squamipinnis has the fastest growth rate and reaches its maximum length (L_∞) earlier. O. lidole has the slowest growth rate. The population of chambo species (O. squampinnis and O. karongae) in Lake Malombe have a slightly higher K (Von Bertalanffy plot) than the same species in south-east arm: growth is 10% – 17% faster.

Table 7.1. Estimated annual growth parameters and growth performance indices.

Figure 7.2 Fishing mortality rates obtained in length cohort analysis of the south-east arm of Lake Malawi chambo species: O. lidole (A), O. squamipinnis (B) and O. karongae (C). For Lake Malombe O. squamipinnis and O. karongae are combined (D).

7.2 ESTIMATION OF MORTALITY RATES

Total mortality rate (Z) was estimated using the linearized length-converted catch curve method, using the growth parameters from the Von Bertalanffy plot as input data.

Natural mortality rate (M) usually is a difficult parameter to estimate in the absence of an unexploited resource. In this investigation the Rikhter and Efanov (1976) method was used to obtain M values for the younger age groups and a sensitivity analysis was carried out to obtain M values for the older and oldest age groups. A value of M = 0.4 was used for the 0–12cm size group (one year old fish), M = 0.3 for 12–20cm group (two year old fish) and M = 0.2 for >20cm group (three year and older fish).

The results of the catch curve analysis are given in Figure 7.1 and Table 7.2. Total mortality rate (Z) is slightly higher in O. squamipinnis than the other two species in the south-east arm, but Z is much higher for the Lake Malombe chambo species.

Table 7.2. Estimated total mortality rates (Z)

During its life cycle, chambo are exploited by different fisheries in a sequential order (Figure 7.1): small-meshed seine net fisheries, in particular kambuzi seines and nkacha nets, exploit the juvenile stages (0+, 1+ years of age) in the shallow near-shore waters, while the large meshed gillnet fisheries exploit the older fish (3+ and older fish) that have moved into deeper waters. The 2+ fish apparently escape heavy exploitation.

The distribution of the fishing mortality (F) per length group, from cohort analysis, is given in Figure 7.2. In the south-east arm, maximum fishing mortality (F_max) is at 36cm, 28cm and 26cm for O. lidole, O. squamipinnis and O. karongae respectively. In Lake Malombe two distinct peaks of fishing mortlity are observed; one at 10cm due to kambuzi seines and nkacha nets and one for the adult chambo caused mainly by the gillnet fisheries. In the south-east arm the fishing mortality of juvenile chambo is seemingly not very high. This is possibly due to the small number of nkacha nets in use in the area and the fact that seine nets have been inadequately sampled. The analysis also shows that 70% of the fishing mortality of O. lidole is caused by semi-industrial and industrial fisheries. About 53% of the fishing mortality of O. squamipinnis is caused by semi-industrial and industrial fisheries whereas in O. karongae 73% of the fishing mortality is caused by the artisanal fisheries.

7.3 ESTIMATION OF STOCK SIZE AND FISHING MORTALITY FROM COHORT ANALYSIS

Jones's (1984) length cohort analysis was used to estimate stock size and fishing mortalities (F). The fundamental assumptions of this method are that fishing effort and recruitment remain constant, thereby affecting all age classes in the same way, and the length composition data are assumed to represent an average situation. The input parameters are: terminal exploitation rate (F/Z), M, K, L_∞ and a and b in the length-weight relationship. The terminal exploitation rate was chosen so that the exploitation rate for the last 4 length-groups became approximately equal as they are all fully exploited. The values chosen were obtained by iterative trials.

Figure 7.3 South-east arm of lake Malawi. Chambo yield and biomass estimates for a range of effort levels, which are expressed relative to the present (1991) effort level (= 1). O. lidole (A), O. squamipinnis (B) and O. karongae (C).

7.4 PREDICTIVE ANALYSIS OF EQUILIBRIUM YIELD AND STOCK BIOMASS

Yield (catch as weight) and stock biomass were predicted for various levels of fishing effort, using the length-based Thompson and Bell analysis for the south-east arm of Lake Malawi. The method assumes constant recruitment and predicts long-term catches. Input data are fishing mortalities per length group and recruitment in numbers (from cohort analysis). It was necessary to develop a different model for a predictive analysis of the Lake Malombe chambo fishery (see section 10.6), since the stocks are no longer in a steady state (Van Zalinge, et al., 1991).

Chambo prices per kilogram are regularly collected by the MTF catch-effort data collection programme and are also obtained from Maldeco Fishing Company. For the south-east arm of Lake Malawi average 1991 prices were: MK 1.04 per kg for lengths between 3–15cm, and MK 1.92 per kg for lengths between 16–40cm

Table 7.3 and Figure 7.3 give the results of the Thompson and Bell model for a range of effort levels which are expressed relative to the present (1991) effort level (= 1). For O. lidole and O. karongae the maximum sustainable yield (MSY) and maximum sustainable economic yield (MSE) are at 70% of the present effort level. For O. squamipinnis the MSY and MSE are at 60% of the present effort level. The estimated MSY and biomass for the chambo stocks in the south-east arm are 3510t and 9883t respectively.

In Lake Malombe the situation is different. Present recruitment is very low due to a severely depleted parent stock and continuous exploitation of juveniles and spawners.

Table 7.3 Current and predicted yields of the various species of Chambo in the south-east arm of Lake Malawi (Thompson and Bell method).

7.5 F_0.1 (F NOUGHT POINT 0.1) TARGET

This analysis was done only for the chambo stocks of the south-east arm of Lake Malawi (Table 7.4).

The parameter is defined as the level of fishing mortality (F) at which the increase in yield for one unit of effort is one tenth of the increase experienced at very low levels of effort. It is approximated by using 10% of the fishing mortality at the maximum rate of yield-per-recruit. The F_0.1 target lies at lower effort levels than F_msy derived from Surplus Yield, Thompson and Bell or Beverton and Holt models. Thus it allows for improved profitability in the fishery and will not lead so easily to growth, or recruitment overfishing.

Table 7.4 F_0.1 TARGET

7.6 MESH SELECTIVITY

The mesh selectivity of various chambo exploiting gears in the project region was investigated using length-frequency samples from different mesh sizes. For chambo seines, kambuzi seines, nkacha nets, chirimila nets, pair trawls, mid-water trawls, bottom trawls and chambo ringnets, cumulative selection curves (assuming a trawl-type selection) were estimated for each mesh size and the 50% retention point (optimum length) was obtained from the curve. For gillnets, the selectivity of 5 different mesh sizes was determined by taking each pair of successive meshes under the assumption they have the same fishing power. (Experimental gillnet data by courtesy of Mr. D. Tweddle).

The estimated optimum lengths (50%) for different mesh sizes for the various gears that exploit chambo are summarised in Table 7.5 Only a meshsize of 90 mm or larger takes either mature or post-spawning chambo. (Details are given in Seisay et al., 1992b).

Table 7.5 Lengths at which 50 % of Chambo caught is retained by different mesh sizes of the commercial gears.