Title: Evaluation and improvements of a satellite based forecast system for Sri Lankan yellowfin tuna fishery

Final project
Year of publication:
Place of publication:
Number of pages:
Supervisors: Hoskuldur Bjornsson


The relationship between yellowfin tuna (Thunnus albacares) catches in the Northeast Indian Ocean, by Sri Lankan long-line fleets, and oceanographic variables obtained from remote sensing satellites was studied to improve a forecasting system for the fishery. Sea surface temperature (AMSRE, AVHRR), chlorophyll-a concentration (MODIS), sea surface height (TOPEX/Poseidon) were analyzed in relation to catch data expressed as catch per unit of effort (CPUE), which was calculated as the number of fish caught by 100 hooks. Splitting the yellowfin fishing areas into three regions, NW, NE and SW, the spatial and temporal variability of oceanographic parameters were determined in relation to CPUE. An existing fishery forecasting system that was based on satellite data was evaluated with updated fishery and satellite data. The results indicate that highest CPUEs corresponded with areas of SST 28.0–30.0 oC, CHL 0.05–0.4 mg m3, and SSH 200-220 cm during the study period. Slight variations of these parameters were observed with time in different fishing areas considered. The CPUE varies between 0.07–14.0 and the mean CPUE was 1.3 during the year 2008. However, the fishery is highly affected in NW fishing area during SW monsoon while other two areas were not influenced by monsoons. To identify the functional relationships between the environmental variables and CPUE, generalized additive model (GAM) was applied. The areas of highest CPUEs predicted by the model were consistent with the potential habitats on the prediction and observation data. Sea surface temperature and chlorophyll-a concentration were statistically significant within predicted ranges while sea surface height showed a flat relationship. These parameters seem to be important in controlling yellowfin tuna distribution in the region.

Documents and links