Abstract
The selection of spatial scales is of particular importance in modeling relationships between fishery abundance and its influencing factors, because these relationships are significantly affected by spatial scale. Here, we explore the spatial scale effects of catch per unit effort (CPUE)–factor relationships for Ommastrephes bartramii in the northwest Pacific. The original commercial fishery data and oceanographic factors were tessellated to 12 spatial scales from 5′ to 60′ with an interval of 5′. Under the original scale and 12 tessellated scales, we constructed the generalized additive models ( GAM s) to model the relationships between the O. bartramii CPUE and the influencing factors, including Year, Month, Latitude (Lat), Longitude (Lon), sea surface salinity ( SSS ), sea surface temperature ( SST ), sea surface chlorophyll‐a (Chl‐a) concentration, and sea surface height ( SSH ). Our multi‐scale analysis showed that the relationships are sensitive to spatial scales. Among the factors, Year, Month, and SSS share quadratic polynomial scaling relations; Lat, SST , and Chl‐a illustrate power law scaling relations; Lon has a linear scaling relation; and SSH presents an exponential scaling relation. Considering the scale sensitivity of the factor sort‐order and the accumulation of explained residual deviance in GAM , we suggest 30′ – 45′ as the optimal range of spatial scales for analyzing the CPUE –factor relationships for O. bartramii . Our research improves understanding of the impacts of changing scales in fisheries and provides a potential method for the selection of a suitable spatial scale for fisheries analysis and resource surveying.