The estimated production of fish in Malawi is 80,000 tonnes per year. Of this tonnage, 50,000 come from Lake Malawi and 30,000 from other water bodies. It has been estimated that fish provide more than 70% of animal protein consumed by the people of Malawi and are therefore an essential component of the national diet. Also fishery operations are a significant source of income, directly or indirectly, for more than 200,000 people.
The fisheries of Malawi can be divided into the small-scale subsistence sector, the artisanal (traditional) sector, the semi-industrial and the industrial sectors. The semi-industrial and industrial fisheries, using pair- and stern-trawlers, operate in eight areas confined to the southern part of Lake Malawi. The artisanal fisheries operate in all lakes, rivers, lagoons, flood-plains and seasonal streams, using a variety of gears such as gillnets, beach seines, open seines, longlines etc. The fishing vessels are canoes and plank boats with or without small outboard engines.
Data collection from commercial fisheries started in 1968, although some data from the early 1950s are available for the chambo ringnet fishery.
For administrative and data collection purposes, the major fishing waters of Malawi are divided into sampling areas which are further subdivided. These divisions create major and minor sampling strata. The system currently in use to collect fisheries statistics relies on the semi-industrial and industrial fisheries operators submitting, to the Department of Fisheries, catch and effort records on standard forms on a monthly basis. The submission of these records is a condition of the granting of the license to fish. The standard forms contain information such as fishing date, the number of trawl or ringnet hauls and the weight and kind of fish caught each day.
The artisanal fisheries are monitored by a boat-based system developed by Bazigos (1974) and introduced in 1976. The data are obtained by means of annual Frame Surveys (FS) and Catch Assessment Survey (CAS). The frame survey involves counting and recording the number and type of boats and gears on each of the fishing sites along the shoreline. It is conducted every year in August, in all the minor strata, by Technical Assistants assisted by Fish Scouts.
The CAS relies on sampling, unlike the FS. Catch and effort records are collected on 4 randomly selected sampling sites identified in the frame survey in each minor stratum and each site is visited 4 times in a month. At each sampling site, the recorder (Fish Scout) carries out a census of numbers and types of craft and fishing gear found on the selected beach. For the next three days, catch and effort data is collected from the selected fishing units to be sampled. The type of information collected includes:
These data are sent to the District Fisheries Office for processing and summarising on working sheet forms. The forms are then sent to the Fisheries Research Unit for the estimation of total catches and fishing efforts. Previous evaluations of the systems (Alimoso, 1988 and Stamatopoulos, 1989) have indicated several methodological and operational problems, particularly in CAS, which might lead to statistical errors. However, the main problem is that the system does not take into account gear distribution and the way the gears are operated. Consequently there are known cases of significant fishing activities taking place without being recorded. With relatively rare gears which make large catches e.g. chambo seines, there may be severe under- or overestimations (Alimoso and Tweddle in prep).
Calculations using raising factors may cause problems. For example the chirimila nets are operated by two or three boats and their catches are regularly overestimated since the raising factor is based on the ratio of the number of fishing craft in the minor stratum to that on the fishing beach sampled.
The data recording forms used in the FS and the CAS are complex. This complexity leads to the misallocation of the data entered. The manual transfer of data from form to form and the manual calculations which follow have been shown to be responsible for significant errors in the accepted statistics.
In view of the observed inefficiencies and assessment errors which are seen to arise from the current systems, it is considered imperative that the Department of Fisheries should improve the existing fishery statistics collection programme. An attempt has been made to minimise errors in operational aspects by reorganising the data collection and, in addition to take advantage of modern computerised data handling techniques to minimise the errors inherent in manual transcription and collation of the data.
In the light of the shortcomings of the existing systems, the Chambo Fisheries Research Project has developed two fully computerised systems: the Malawi Fisheries Information System (MFIS) and the Malawi Traditional Fisheries (MTF) for the collection of fisheries statistics in the eight minor strata of the south-east arm of Lake Malawi. The design and operation of MFIS and MTF are described by Alimoso et al. (1990) and Stamatopoulos (1989).
MFIS has been designed for storage, processing and reporting statistical information from semi-industrial and industrial fisheries. It is based on the ongoing system but with the routine statistical processing automated. The recording forms used are those of the existing system.
MTF is a gear-based system designed for storage, processing and reporting statistical information from the artisanal fisheries. The system involves the collection of data from a few selected landing sites or beaches and then uses the sample data to estimate the total fish landings by applying raising factors.
The selection of landing sites is based on the results of the annual Frame Surveys and the following points are taken into consideration:
As in the annual frame survey carried out by the Department of Fisheries, the Technical Assistants, who are the field recorders in this project, visit all landing sites along the shoreline and record, among others, the name of the fishing site, the name of the fisherman, number and type of gear and craft owned, (Alimoso et al., 1991). In addition, they record gear characteristics such as headline length, depth of net and mesh size.
Having selected the landing sites and determined the number of landings to be sampled from each gear and from each minor stratum, a sampling survey is planned. The sampling survey involves sampling in space and time, and has two main sampling activities also carried out by Technical Assistants, namely:
Both sampling activities are conducted for each gear type during selected days of the survey period. A survey period is defined by an itinerary given to the beach recorder every month. The itinerary simply indicates beach name, the gears to be sampled and the sampling days chosen at random within a month.
On each sampling day, the recorder carries out a gear census. This is simply done by recording the number and types of fishing gears that exist on the beach and those out fishing on that day.
On the same sampling day, the recorder gathers information on catch and effort from some of the arriving fishing units which are selected and sampled randomly. Besides general information such as name of fisherman, data, beach, gear type etc, a recorder records catches by species and gear in the normal units used by the fishermen, e.g. tins or crates. In cases where catches are reported in dozens, the amount would be converted into weight in kilograms by multiplying by the mean weight of one dozen before entering on the form. At the same time fishing effort relating to the catch is also recorded as the number of seine hauls or gillnet sets. Catch and effort data pertaining to the day previous to the day of actual sampling is also obtained by interviewing the fishermen. The fishermen are asked about the prices obtained for the catches for both days.
At the end of the survey, estimation of monthly catch and fishing effort for each gear type and for each minor stratum proceeds as follows:
All the catch figures collected during the month are converted into weights in kilograms by applying predetermined conversion factors. These are summed to provide a figure for the total catch. Similarly, all the effort figures are added together to provide a total monthly effort (f). An estimate of the monthly catch-per-unit (CPUE) is calculated as:
CPUE = C/f
The Gear Activity (GA) is estimated by dividing the total number of gears that were found to be fishing at the time of sampling by the total number known to exist at the landing sites. Therefore, total monthly fishing effort (E), in the minor strata is estimated by:
E = GA × D × M
where M is the total number of fishing gear units in the minor strata that were counted during frame survey and D is the number of fishing days in a month.
The estimated total monthly catch (Y) for a particular gear in the minor stratum would, therefore, be given by:
Y = CPUE × E
In addition to the above information, two sampling programmes, for length-frequency analyses and cost and earnings investigations, were fully integrated with the work of the survey.
The length-frequency programme involved collection of data on chambo size-ranges being landed in a given area from artisanal, semi-industrial and industrial fisheries and for each gear-type. Furthermore, biological information such as fish length, weight, gonadal state, sex, age etc. were also collected from length-frequency samples.
In order to assess the economic viability of the operating units, a cost and earnings survey was carried out. Cost and earnings data were gathered by interviewing fishermen for information on the selling prices of their catch and their expenses in connection with their fishing operations. The information collected included the amount of cash payments to crew members and the cost of fishing operations such as replacement of warps, supply of paraffin, food etc.
All the above information, except the length-frequency data was then entered directly into MTF field forms and all analyses were done by computer using these forms as source documents. The length-frequency data was analysed using the Length-based Fish Stock Assessment (LFSA) computer programme (Sparre, 1987).
A comparison of the catch and effort estimates produced by the gear-based sampling system (MTF) and the boat-based CAS, was undertaken by Turner et al., (1992b). From this study it was found that MTF provides much more reliable estimates, showing smooth trends in gear activity and catches within the year, although total production estimates for all species caught and all gears used differed little between the systems in areas of high fishing activity.
Since the design and operational aspects of MTF system mitigates many of the problems identified in the FS and CAS data collection systems, collects the same information and does not require changes in the existing procedures or the mobilization of human resources, the recommendation made by the project to the Department of Fisheries that the system should be adopted on a nationwide basis without delay, was accepted. This will ensure the collection of better quality of fisheries statistics which is the central element in monitoring and managing the fish resources of Malawi.