FIRST UGASTAT CAS RESULTS FOR LAKES VICTORIA, ALBERT AND THE EDWARD/GEORGE-KAZINGA CHANNEL COMPLEX.

by

UGA/87/007 Biostat Group

edited by

E.J. COENEN

1. INTRODUCTION

In order to make the UGASTAT software package (for the processing of Catch Assessment and Frame Survey data) operational and because Mr. Coppola, computer consultant, was unable to come to Uganda for a third visit, a study/training tour was organized in July 1991 to the Statistics Division of FAO, Fisheries Headquarters in Rome. This arrangement gave an opportunity to two biologists/ statisticians to test UGASTAT with real data; to make the CAS analysis part of the UGASTAT statistical software operational; to start the draft of the UGASTAT reference manual and to select several FAO publications for the newly rehabilitated Fisheries/Game Departments Library (Tumwebaze & Ikwaput, 1991; Coppola, Coenen, Ikwaput and Tumwebaze, 1991).

A brief description of the new CAS methodology, background, design work, testing, etc. is given in Chapter 2.

In August 1991, the first CAS data were processed at the project headquarters in Entebbe. A brief review and discussion of the results is given in Chapter 3 and monthly CAS summary results per waterbody are presented in Appendix A. In Appendix B, the characteristics for all the landings of Lake Victoria and Albert (based on the Frame survey results) are presented for each minor stratum sampled.

2. THE NEW CATCH ASSESSMENT SURVEY DESIGN

The main criteria for the improved or new methodology for the Catch Assessment Surveys were the following: complete coverage of the area; feasability of the survey in terms of manpower and means available; harmonization and standardization of methodology and survey forms to be used; sampling at random of fish landings and boats to survey; final estimates of catch and effort to be accurate and precise; major and minor strata borders to coincide with administrative borders of districts, counties, etc.; compatibility of the survey forms with the data entries needed by the UGASTAT CAS computer programme.

A first trial with new forms started in the Lake George/Edward complex in May 1990. Three forms were used : a daily fish landing sheet, a monthly landing sheet summarizing the daily totals and a monthly water body sheet summarizing the monthly fish landing totals.

During the visits of the computer consultant and the consultant in CAS design, the methodology and procedures were gradually improved to comply with the criteria set (see above) and were finalized in May 1991 (Coppola, 1991; Malvestuto, 1990). The final CAS design can be summarized as follows:

• each water body is divided into major strata corresponding to ecological zones and into minor strata mainly corresponding to feasable sampling areas in terms of island/mainland, manpower availability, logistics, administrative boundaries, etc. A codification system was developed assigning specific codes to waterbodies, strata, substrata, districts, counties, subcounties, landing sites, etc;

• sampling of landings within each substratum is done according to a fixed sampling scheme. Landings are chosen at random but with a probability equal to the importance of the landing in terms of active fishing canoes (a bigger landing has more chance to be chosen than a small one);

• chosen landings within a substratum are sampled according to a predetermined time schedule varying between every second and every 4th day;

• surveying of boats landing is done by complete enumeration at smaller fish landings or by sampling every nth boat at bigger landings. Per species, the catch is weighed, counted (optional) and valued;

• the number of survey forms is reduced to two. The amount of calculations to be done by field staff is limited to a strict minimum of only adding totals and calculating two raising factors. The two standard forms in use are the daily fish landing sheet CAS1 and the monthly substratum sheet CAS2 containing the CAS1 totals;

• the major part of the calculations is done by the UGASTAT CAS processing programme. Only the CAS2 codes, totals and raising factors are entered into the computer. All calculations and catch/effort estimates are processed on a monthly basis by the computer. The final output is a monthly printed computer CAS report for each water body, per minor stratum. The analysis part per district still has to be completed and also the final printed output needs a lot of refinement. At present, it takes an extra man/day of manual calculations to turn the computer output of CAS results for one month and one waterbody into an presentable summary table.

In brief, the main characteristics of the new CAS methodology for the project water bodies can be summarized as follows:

The CAS design, as described above, was implemented for Lake Victoria as from March 1991 (after a trial period during February 1991 in several mainland and island substrata in the Entebbe Region); for Lake Albert as from June 1991 and for the Lakes Edward/George and Kazinga complex as from July 1991. Before the start of each new CAS, refresher courses were held at regional level for all minor strata supervisors.

The new CAS methodology has to provide reliable catch and effort estimates for the project water bodies, except for the Lakes Edward/George/Kazinga complex where only the legal activities can be assessed. It has proven impossible to assess correctly the number of illegal canoes and landings. Moreover, despite the fact that legal law enforcement exercises are mounted on a continuous basis whereby hundreds of illegal canoes are destroyed monthly, little effect has been obtained in reducing the illegal fishing activities. The project therefore tries to estimate the extent of illegal activities by extrapolating (using a ‘illegal activity raising factor’) the catch/effort estimates for the legal activities up to a level that the final estimate reflects a catch/effort assessment for both the legal and illegal fishing activities.

3. PRELIMINARY CATCH ASSESSMENT SURVEY RESULTS USING UGASTAT

The first CAS results, using UGASTAT, were obtained at the end of August 1991, after a study tour of two project headquarters biologists/statisticians to the Statistics Division in FAO (Rome) where UGASTAT was updated with the assistance of the FAO computer consultant (Tumwebaze & Ikwaput, 1991).

CAS results for Lake Victoria for the period March-May 1991 reveal a total catch of 32,462.5 tonnes, giving a monthly average of 10,820.8 tonnes and an estimated, extrapolated annual catch of 129,850 tonnes (as compared to the UFD figure of 119,900 tonnes for 1990). A striking result is the fact that almost 62 % of the catch is recorded in the island and not the mainland landings. Three species, Lates niloticus (60.8 %), the tilapiines (21.6 %) and Rastrineobola argentea (14.7 %), constitute 97.1 % of the total catch. The other main genera in the catch are Clarias, Protopterus, Bagrus and Mormyrus. Less important ones are Barbus, Labeo, Schilbe, Haplochromis and Alestes. The average landing price per kg of fish increased from 153 in March 1991 to 213 Ug.Shs in May 1991. The cheapest fish is Rastrineobola argentea (131 Shs in May 1991), the most expensive one is Bagrus (310 Shs in May 1991). An average, monthly landing value of 1.96 billion Ug.Shs or an estimated annual landing value of 23.58 billion Ug.Shs was recorded during that period.

CAS results for Lake Albert for June 1991 reveal a total catch of 2,068.1 tonnes or an estimated annual total of around 24,800 tonnes (as compared to the 1990 UFD figure of 19,500 tonnes). Five genera, Hydrocynus (32.2 %), Lates (31.5 %) Alestes (10.9 %), the tilapiines (10.9 %) and Bagrus (6.3 %) constitute 91.8 % of the total catch. Less important genera are Barbus, Labeo, Clarias, Protopterus, Synodontis, Auchenoglanis, Distichodus, Mormyrus. The average landing price in June 1991 was 134 Ug.Shs, the cheapest fish being Lates (88 Shs) and the most expensive one Distichodus (238 Shs). A monthly landing value of 277 million Ug.Shs or an estimated annual landing value of 3.3 billion Ug.Shs was recorded.

CAS results for Lake Edward during the period January-April 1991 reveal an average, monthly catch of 299 tonnes or an estimated annual catch of 3,588 tonnes. CAS results for the Kazinga Channel landings for April 1991 give a monthly total catch of 24.4 tonnes or an estimated annual catch of 292.8 tonnes. CAS results for Lake George for the period May 1990–May 1991 indicate an average, monthly catch of 196.7 tonnes or an annual catch of 2,360 tonnes. For the whole Lakes Edward/George and Kazinga Channel Complex, the estimated annual catch for the legal (recorded) canoes would be 6,241 tonnes. Using an “illegal catch raising factor” of 2 to 3 would bring the actual, annual catch between 12,500 and 18,700 tonnes (as compared to the 1990 UFD figure of 5,500 tonnes). Major genera caught are the tilapiines, Protopterus, Bagrus and Clarias. Average landing prices per kg, in April 1991, varied between 274 and 315 Ug.Shs.

4. CONCLUSIONS/RECOMMENDATIONS

Sampling frames for improved CAS were developed, tested and implemented, bearing in mind standardisation and cost effectiveness (low survey-cost, high precision and accuracy of final yield estimates). At FSISU Headquarters, computers were installed and a special software package UGASTAT developed for the processing of Frame and Catch Assessment Survey data.

The first CAS results generated indicate realistic catch magnitudes in the order of previous UFD figures. The UGASTAT software, however, still needs to be further tested, adjusted and refined to improve on the final outputs printed. Ideally, a complete cycle of one year of CAS results should be processed by UGASTAT before a final evaluation of the efficiency of the new CAS design and the UGASTAT package can be made.

In order to continue the activities initiated, and apart from possible donor assistance in the future, the Government of Uganda will have to avail an adequate budget in order to sustain future field and FSISU headquarters activities in fishery statistics data collection and processing as designed and implemented by the project.

Appendix A

LAKE VICTORIA

CAS. MARCH 1991

Others : Barbus spp.: Labeo spp.: Schilbe spp.: Haplo. spp.: Alestes spp.

LAKE VICTORIA

CAS. APRIL 1991

Others : Barbus spp.: Labeo spp.: Schilbe spp.: Haplo. spp.: Alestes spp.

LAKE VICTORIA

CAS, MAY 1991

Others : Barbus spp.: Labeo spp.: Schilbe spp.: Haplo. spp.: Alestes spp.

LAKE ALBERT

CAS, JUNE 1991

Source : Project UGA/87/007 Fisheries Department