BIOSTAT REPORT NO. 12

The Present Status of Fisheries Data Collection and Analysis In Uganda

by

J. Ikwaput & R. Tumwebaze

Octoberr, 1989

1. Introduction :

Collecting fisheries (Landing and Marketing) statistics is one of the the duties of the Assistant Fisheries Development Officers and the Fisheries Assistants who are based at fish landings and/or rural or town markets. These data, compiled or not, are submitted to the Regional and District Fisheries Offices, who in turn, after checking and re-compiling, forward them to the Statistical Unit at the Fisheries Department Headquarters at Entebbe for final analysis before utilisation.

2. Fisheries Data Collection at Landings :

Fisheries Field Officers at the landings are expected to collect daily fish landing statistics and fish marketing statistics. Where two or more landings are supervised by one officer, a sampling system is adopted. The officers record the total number of boats fishing each day, the number sampled, and the total number and weight of fish landed and the value obtained. Sampling systems in use vary according to the size of the landing, the number of boats present and the amount of catch landed at the landing. In small landings, with a few boats and less catch, total enumeration of catch of all the boats is done. If a weighing scale is available all the fish is weighed. When there is no scale, the average weight is fixed according to the mesh size used and the fish size based on the experience of the officer. This average weight is then multiplied by the total number of fish counted to obtain the estimated weight for the day. In big landings with many boats, sampling of boats is done. However, sampling methods differ. The officer can decide to sample say the first boat then to sample every “i”th boat systematically depending on the number of boats at the landing. If the catch is small and there is a weighing scale available, the fish in each sampled boat is sorted according to species, counted and then weighed. Where there is no scale, the fish is counted and the average weight which is fixed according to mesh size of the nets used and fish size is determined and multiplied by the total number of fish per species in the catch. When the catch is big, fish are normally tied in bundles by the fishermen according to species and sizes (Fig. 1). The officer counts the bundles in the boat as they are sold noting down the price, and then a sample of bundles is picked, weighed, and the average weight per bundle is then multiplied by the total number of bundles in the boat to obtain the total weight of the boat load. In case of big Nile Perch which are not bundled, a count is made and an average weight of a sample obtained. This is then multiplied by the total number of the fish in the boat. At the end of the day the average weight per boat for the sampled boats is obtained and this is raised for the total number of boats at the landing to give an estimate for the day's catch.

Some landings receive both fresh and processed fish. Where processed fish is landed, the number of bundles (tied as in fig. 2) are counted. These bundles are made of two pieces of wooden rings, polythene sheeting and ropes. If bundles are small, some are weighed, and their average multiplied by the total number of bundles. In the case of big bundles, the number of pieces per bundle is obtained by interviewing the fisherman. Pieces are then removed from a sample bundle, weighed and an average obtained which average is then multiplied by the total number of pieces in the bundles to give the total weight landed. Where no weighing scale is available, the following estimate is made: for Tilapia, the specimens are usually packed in bundles which contain an average of 250–500 pieces according to size, and weighs about 80–150 kgs. So the total weight is equal to 80 – 150 times the number of bundles. For Nile Perch, bundles contain usually an average of 100–200 pieces and weighs about 120– 250 kgs. This species is usually cut into two pieces when smoking, so the number of fish is known by dividing the number of pieces by two.

On landings where processing of fish takes place, the amount smoked, salted, fried or sun-dried may be recorded daily or weekly depending on the method used and the quantity processed. The number is obtained from fish processors and for weight the officer weighs a few pieces, gets the average weight and multiplies it by the number of pieces processed. However, not all officers do record the fish processed at the landings. (Field Officers, Pers. Comm.)

3. Fisheries Data Collection at Markets :

In each Region and District, some officers have been assigned to collect marketing data (weight and value) of fish sold in each market. But in practice, not all markets are sampled. Unlike landings which operate on daily basis, markets vary in the number of days in which they operate. Big markets like Busia fish market (Tororo District) and the markets in Kampala operate on a daily basis while the small markets operate once or twice in a week. Generally, all big markets are supervised by fisheries officers. They record daily weights of fresh and processed fish brought into the market, the average price per kg. and the value obtained. Fresh fish is supplied to the market on daily basis while the processed fish is usually marketed during some specific days. Most of these big markets have their fish delivered by pick-ups. The officer collects the data at the time the fish is being delivered to the market. They sample the first vehicle to arrive and determine the sequence in which to sample the rest of the vehicles. Fresh fish is removed from the vehicle, loaded on a wheelbarrow according to species, the number of wheelbarrow loads are counted, a sample of fish is removed from a wheelbarrow and weighed to find average weight per fish. The number of fish on the wheelbarrow is then counted and the total multiplied by the average weight to obtain the total weight of a wheelbarrow load. The weight of fish in one vehicle is then derived from the weight per wheelbarrow times the number of wheelbarrows off-loaded. At the end of the day, the officer gets the total weight by multiplying the average pick-up load with the number of vehicles which operated that day. For processed fish the number of bundles is counted, pieces are removed from a sample bundle, weighed and an average obtained which average is then multiplied with the total number of pieces in the bundles to obtain the total weight brought into the market. The number of pieces in the bundles is obtained by interviewing the fishmongers (Tumwebaze, 1989).

For small markets, the officer may collect data on all the market days or he may just sample once or twice a month. These sample totals are either raised to the total number of market days in the month or they are just submitted as such to the Regional Fisheries Office or District Fisheries Office with or without indication of what they represent.

During submission, processed fish data are not categorised under sun-dried, salted and smoked except in a few regions.

4. Submission of Data from Landings or Markets to the Regional or District Fisheries Office (RFO/DFO):

At the end of the month, the collected data are compiled and forwarded by each officer at the landing or market to the RFO/DFO. However, the state of the data submitted to the RFO/DFO is not uniform for all landings and markets. Some officers forward their daily raw data to the RFO/DFO while some compile it and apply raising factors to give an estimate for the whole month before submission is done. Where compilation is done, some officers do not indicate whether an attempt has been made to estimate for the whole month, for all the boats and all landings administered or if it is only the total for the samples taken. The same applies to the markets. This marks the beginning of a recurrent problem of inaccuracy of some data which continues up to the figures submitted to the headquarters.

5. Submission of Data from Regional and District Offices to Fisheries Department Headquarters:

At the RFO/DFO, it is the responsibility of the Regional Officer or District Officer to check on any errors and omissions and to make estimates for the landings and markets which submitted only their raw data. The compiled data are then forwarded to the Headquarters (as shown in attached table 1). However, the RFO and DFO may find it difficult to detect any errors and to estimate the monthly production for a particular landing or market if no indication is made on whether the submitted data is a complete estimate for 30 days or not, how many days were sampled, how many boats were sampled, how many are present at the landing, and how many market days are in a month. The submission is done on monthly basis and it reflects species composition and monthly weights and values for each landing and market. The figures normally submitted by RFO/DFO only reflect the data for the sampled landings and markets and not for all landings and markets in the region (since the exact number and magnitude of the landings and markets in the region is not known and thus no raising factor can be applied).

6. Compilation and Analysis Procedures of Data at Fisheries Department Headquarters :

At the fisheries headquarters, the statistics and information unit is responsible for extracting the figures from the monthly RFO and DFO reports. The unit recompiles the data, appraising their quality and identifies errors and omissions. The first stage involves tabulation of monthly fish production for each submitted landing showing monthly species total weight and value (see table 2). At the end of the year, total yearly estimated production is made for each sampled landing. These annual estimates are derived from the average monthly weights for the submitted months raised to twelve months. From this a second tabulation is made for all sampled landings per fisheries region showing annual weight and value for each species.

Second stage involves tabulation of all the sampled landings for a given water body showing annual estimated weight and value per species (see table 3). These are added up to give the total annual catch for each species for the sampled landings. The total annual catch figure per species for all species combined is then raised upward by dividing it by the total number of boats in the sampled landings and multiplying the average boat catch by the total estimated number of boats in the water body to give the total annual estimated catches and values for the whole water body and species totals. However, it is not always easy to make the estimate for the whole water body because at times the total number of boats is not known and available data from frame surveys is usually very old.

Third stage involves tabulating the total for all the water bodies to give a comparison in total catch between lakes, rivers, ponds, etc. and to derive the total annual catch for the whole country (see table 4).

Marketing statistics are also compiled in stages. First stage involves compiling marketing figures for each submitted market indicating the form in which fish is marketed i.e. fresh or processed (e.g. sun-dried, salted, smoked or fried) giving quantities of each form and the retail price per species. Weights for processed fish are not converted to fresh weight. Estimates for each market are then made at the end of the year by raising the submitted totals to 12 months (Table 5). Second stage involves compiling together the sampled markets in each district giving annual estimated marketing totals for each sampled market and indicating the form in which fish was marketed. Annual totals for the sampled markets in the district are added up (Table 6). Information on the source of fish supplies is sometimes given. (Ikwaput, pers. obs.)

7. Discussion:

Data collected in some field stations are not complemented by additional information on fishing and fishing gear characteristics (gear specifications, number of boats, etc.). Also, some Regional Offices, when submitting the data to the Headquarters, do not indicate whether it is an estimate for the whole month or just the totals for the sampled days only. Mostly, no attempt is made to make estimates for the unsampled landings which the officer may have some knowledge on. Total number of landings in the lake regions are not indicated because most officers do not know the total number of landings in their regions. These short-comings make it difficult to use a defined formula to estimate annual catch for a certain water body. On the other hand, for Lakes Edward/George and the Kazinga channel for example, where landings are gazetted, the number of landings is well known. But there exist quite some illegal (unlicensed) boats on which the officers do not report. This often results in underestimates of total catches since only the legal boats are surveyed.

Generally, the inadequate submissions of monthly reports from almost all the regions, which are often late and based on incomplete and fragmentary returns, hamper proper compilation and analysis of figures for estimating accurate total catch figures.

Compilation of market data is more difficult than that of landing data. Like catch figures from landings, marketing figures from markets are also not uniform in their format of submission. Some field staff only give weights and values for marketed fish indicating form of marketing as fresh or processed. No attempt is made to break down the processed fish according to form of processing, i.e. sun-dried, smoked, salted and fried. Some field staff do indicate the source of the fish while others do not. During re-compilation and analysis at the Fisheries Headquarters, some problems arise. Where no indication is made on whether the submitted totals are only sampled totals or estimates for the whole month, it becomes difficult to make a decision on whether to apply a raising factor or not. Even when the sampled days are given without indication of the total number of market days in the month, it is difficult to make any estimates since some markets do not operate on daily basis. Also number and magnitude of non sampled markets is not known and there has never been any frame survey on markets. This generally makes it difficult to estimate correctly how much fish is marketed in each district (Ikwaput & Tumwebaze, Pers. Obs.).

8. Recommendations:

1.     A frame survey should be carried out to ascertain the exact number of landings and to find out their magnitude in terms of active fishing boats. Also, a frame survey should be carried out for markets to find out their numbers in each district, the days on which they operate and how many times a month.

2.     A uniform format (standardised forms) for tabulation of data collected from the field should be developed, to reflect: number of boats, gear specifications for each landing, and number of landings in each region with an indication of both sampled and unsampled ones and the raising factors to apply. Also a uniform format for collecting marketing statistics should be developed plus a conversion table to be used to convert processed weight for each species to fresh weight.

3.     Where data from small landings are lumped together with that of big landings, footnotes should be included on the number and/or names of landings which have contributed to the total figure under the major landings submitted. Lumping together of data should be discouraged for both landings and species.

4.     There should be a routine system of on-the-spot checks and quality check surveys based on the number and time of sampling days. Sampling schemes for the next month should be indicated in the report submitted by field officers.

5.     A better and regular communication system between landings/markets and Regional/District Offices on the one side and between Regional/District Offices and Fisheries Department Headquarters on the other side should be established to allow for an adequate and timely delivery of reports from the field.

9. Acknowledgement.

We are greatly indebted to Mr. E. Coenen, CTA, project UGA/87/007, and Dr. F.L. Orach-meza, Deputy Commissioner for Fisheries who read through the draft and made many valuable corrections. We would like to register our sincere thanks to our fellow Biologists/Statisticians, Mr. J. Wadanya and Mr. D. Nyeko for their valuable contribution towards this paper.

References :

FIG. 1.

SKETCH DRAWING OF A BUNDLE OF FRESH FISH.

FIG. 2.

SKETCH DRAWING OF A BUNDLE OF PROCESSED FISH.

TABLE 1. FISH CATCHES AND VALUES DURING THE MONTH OF DECEMBER, 1988. - LAKES EDWARD/GEORGE AND KAZINGA CHANNEL - KICHWAMBA REGION (WT. IN Kg., VAL. IN U. Shs.).

SPECIES

Source: Uganda Fisheries Department (1989).

TABLE 2. ESTIMATED FISH CATCHES AND VALUES FROM LAMBU LANDING - MASAKA REGION - 1988 (WT. IN TONS., VAL. IN MILLION U. SHs.)

SPECIES

Source: Uganda Fisheries Department (1989).

TABLE 3. ESTIMATED ANNUAL FISH CATCHES AND VALUES BY LANDINGS FOR LAKES EDWARD/GEORGE AND KAZINGA CHANNEL (KICHWAMBA REGION) - 1988 (WT. IN TONS, VAL. IN M. U.Shs)

SPECIES

Source : Uganda Fisheries Department (1989).