This study investigated the cause of long-term variations in the red sea bream catch off Miyazaki Prefecture, which has the Hyuga-Nada on the Pacific coast as its fishing ground. We analyzed approximately 60 years of fisheries catch data and oceanographic conditions. The red sea bream catch in Miyazaki Prefecture showed negative anomalies in the 1950s, 1970s, and late 2000s and positive anomalies in the 1960s, 1980s, and early 2000s. The catch increased in the same year, or after several years, of synchronization with a decreased sea level at Hososhima, which was related to the lower bottom-water temperature. Two mechanisms were hypothesized to increase the red sea bream catch in Miyazaki Prefecture. In the first process, the catch increases due to the high condition of the fish and a subsequent high reproduction rate as the feeding environment improves following high primary productivity by nutrient-rich cold bottom-water in Hyuga-Nada. In the second process, the catch in the eastern part of the Bungo Channel increases due to an improved feeding environment following high primary productivity by the intrusion of nutrient-rich cold bottom-water from the Pacific Ocean into the Bungo Channel. This leads to an increase in the catch in the northern part of Miyazaki Prefecture two years later due to the southward migration or wide distribution from the eastern Bungo Channel.

For better understanding of hydrographic conditions and improving the accuracy of coastal ocean prediction models, we need to construct high-resolution observation system. In this study, we developed a new CTD sensor that is affordable, robust, and handy for coastal fishermen. We then set up a hydrographic observation system supported by the fishermen and conducted demonstration experiments in the real fishing grounds. This observation system allowed us to collect numerous CTD data at unprecedented spatiotemporal scales in the coastal waters northwest of Kyushu Island. The fluctuation of salinity at 10–20 days’ cycle is found in the Iki Channel from July to September 2019. Involving fishermen has been shown to be very effective for the coastal observation system since the number of fishermen observation data is an order of higher than the existing observation system. The CTD data from the fishing vessels have been assimilated into a coastal ocean model since January 2019, contributing to the accurate prediction of coastal fishing environment.