Abstract: Based on the data collected by scientific observers onboard Chinese Antarctic krill fishing vessel, the present study analyzed the effect of month, bottom depth and fishing depth on catch per unit effort (CPUE) of Antarctic krill fishery operated in the South Georgia Island during the austral winter 2013-2016 using geographically weighted regression (GWR) model and further revealed the drivers of the spatialtemporal differences. The results indicated that higher CPUEs were distributed in the first 10 days of July to the last 10 days of August. The highest mean CPUE was 20.5±21.1 t·h^-1 in 2014, and the lowest mean CPUE was 10.4±10.5 t·h^-1 in 2015. A significant difference was also found in average CPUE values among months. Fishing ground was concentrated in 53°00′S-54°30′ S, 35°15′W-38°30′W. The goodness of fit for GWR model was the highest in 2015, but the lowest in 2016. The spatial effect (positive correlation) of fishing depth on CPUE of Antarctic krill fishery was strengthened increasingly from south to north, but no consistent trend was observed for the spatial effect of bottom depth on CPUE. The goodness of fit for GWR model was higher than that of ordinary least square (OLS) model, indicating that GWR model was more suitable to present the spatial effect of drivers on CPUE. The results derived from the present study can provide effective references for studying forming mechanism of fishing ground and fishery management of Antarctic krill fishery in the South Georgia Island.

Key words: mitochondrial DNA, genetic structure, COII gene, geographic population, genetic diversity., Harmonia axyridis