Species Targeting Information Used to Standardize CPUE for South Pacific Albacore

South Pacific albacore are targeted by the longline fishery operating out of American Samoa and are also caught in other longline fisheries, including those by distant-water fleets of Japan, Korea and Taiwan that target other tuna species. In the MULTIFAN-CL stock assessment of South Pacific albacore, catch and effort data from the longline fleets are used to compute the principal indices of relative abundance for the albacore stock. However, there have been temporal changes in the catchability of the longline fisheries, primarily in the distant-water fleets, as a result of changes in the species targeted. PIFSC scientist Keith Bigelow and colleague Simon Hoyle of the Secretariat of the Pacific Community have computed new indices of South Pacific albacore abundance that account for effects of species targeting. They presented their results at the August 2012 meeting of the Western and Central Pacific Fisheries Commission (WCPFC) Scientific Committee in Busan, Korea.

Figure 1.  Spatial distribution of South Pacific longline fishing effort for 3 clusters over the period 1960 to 2011.  Cluster 1: albacore and yellowfin tuna; Cluster 2: albacore; Cluster 3: yellowfin and bigeye tuna. In general, there is some spatial separation in the clusters, with Cluster 3 predominating at low latitudes, Cluster 1 at mid-latitudes, and Cluster 2 at high latitudes.  In the CPUE standardization analysis, the South Pacific is divided into 4 regions (Region 1 = NW quadrant, Region 2 = NE quadrant, Region 3 = SW quadrant, Region 4 = SE quadrant). The GLM analysis was based on longline sets targeting South Pacific albacore, defined as those in Clusters 1 and 2; Cluster 3 data were not used. Figure 1.  Spatial distribution of South Pacific longline fishing effort for 3 clusters over the period 1960 to 2011.  Cluster 1: albacore and yellowfin tuna; Cluster 2: albacore; Cluster 3: yellowfin and bigeye tuna. In general, there is some spatial separation in the clusters, with Cluster 3 predominating at low latitudes, Cluster 1 at mid-latitudes, and Cluster 2 at high latitudes.  In the CPUE standardization analysis, the South Pacific is divided into 4 regions (Region 1 = NW quadrant, Region 2 = NE quadrant, Region 3 = SW quadrant, Region 4 = SE quadrant). The GLM analysis was based on longline sets targeting South Pacific albacore, defined as those in Clusters 1 and 2; Cluster 3 data were not used. Figure 1.  Spatial distribution of South Pacific longline fishing effort for 3 clusters over the period 1960 to 2011.  Cluster 1: albacore and yellowfin tuna; Cluster 2: albacore; Cluster 3: yellowfin and bigeye tuna. In general, there is some spatial separation in the clusters, with Cluster 3 predominating at low latitudes, Cluster 1 at mid-latitudes, and Cluster 2 at high latitudes.  In the CPUE standardization analysis, the South Pacific is divided into 4 regions (Region 1 = NW quadrant, Region 2 = NE quadrant, Region 3 = SW quadrant, Region 4 = SE quadrant). The GLM analysis was based on longline sets targeting South Pacific albacore, defined as those in Clusters 1 and 2; Cluster 3 data were not used.
Figure 1. Spatial distribution of South Pacific longline fishing effort for 3 clusters over the period 1960 to 2011. Cluster 1: albacore and yellowfin tuna; Cluster 2: albacore; Cluster 3: yellowfin and bigeye tuna. In general, there is some spatial separation in the clusters, with Cluster 3 predominating at low latitudes, Cluster 1 at mid-latitudes, and Cluster 2 at high latitudes. In the CPUE standardization analysis, the South Pacific is divided into 4 regions (Region 1 = NW quadrant, Region 2 = NE quadrant, Region 3 = SW quadrant, Region 4 = SE quadrant). The GLM analysis was based on longline sets targeting South Pacific albacore, defined as those in Clusters 1 and 2; Cluster 3 data were not used.

In earlier assessments of South Pacific albacore stock, circa 2005, analysts used catch-per-unit effort (CPUE) abundance indices based on nominal operational level data (logsheet data of daily fishing activities) for domestic longline fisheries and data aggregated by month and areas of 5° latitude x 5° longitude for distant-water longline fisheries. Since 2008, however, there has been a progression towards developing standardized CPUE indices derived from operational data and Generalized Linear Models (GLMs). The data are spatially stratified into 4 regions south of the equator and between longitudes of 140°E and 250°W, with quadrants delineated by boundaries along 25°S and 180° (Figure 1).The objective of the study by Bigelow and Hoyle was to produce a single standardized CPUE index for each of the 4 regions for longline vessels targeting South Pacific albacore. The study improved on previous standardizations (which included only data from distant-water fleets of Japan, Taiwan and Korea) in several ways: 1) all operational data for longline fishing south of the equator were used from all longline fleets, including the fleet based in American Samoa; 2) trips targeting South Pacific albacore were identified by three alternative types of cluster analysis; and 3) the GLMs considered a negative binomial distribution instead of the lognormal distribution.

Clustering of data was performed based on the proportions of albacore, yellowfin tuna, and bigeye tuna caught during each longline fishing trip. Dendrograms produced in the cluster analysis indicated 3 primary clusters for each region: Cluster 1 represents targeting of albacore and yellowfin tuna, Cluster 2 represents targeting of albacore, and Cluster 3 represents targeting of bigeye and yellowfin tuna. Figure 1 illustrates the spatial variability of longline effort for the 3 clusters over the entire time series. In general there is some spatial separation in clusters, with bigeye and yellowfin tuna (Cluster 3) predominating at low-latitudes, albacore and yellowfin tuna (Cluster 1) at mid-latitudes and albacore (Cluster 2) at high-latitudes. Cluster 3 made up a larger percentage of fishing trips (range=23.0-27.6%) in the northern regions (1 and 2) than in the southern regions (3 and 4, range=6.0-13.2%). The bigeye and yellowfin cluster made up a high proportion in Region 1 until 1980 and a moderate proportion in Region 2 throughout the time-series. Clusters 1 and 2 dominated regions 3 and 4 throughout the time-series.

GLM analyses were conducted to produce time series of standardized South Pacific albacore CPUE in each of the 4 regions, using the data from Cluster 1 and Cluster 2. Figure 2 shows the standardized CPUE in each region along with the nominal CPUE. A rapid decline in standardized CPUE is seen from the early 1960s until 1975 followed by a slower decline thereafter. In Regions 2-4, an increase in the standardized CPUE in the late 1990s is evident. Standardized CPUE declined in 2003 in most regions, and thereafter was stable in the east (regions 2 and 4) and increased in the west (regions 1 and 3).

Figure 2.  Nominal and standardized CPUE for South Pacific albacore by region.
Figure 2. Nominal and standardized CPUE for South Pacific albacore by region.