STANDARDIZATION OF CPUE OF THE NORTHERN‐CENTRAL PERUVIAN ANCHOVY (Engraulis ringens) PURSE SEINE FLEET Dante Espinoza‐Morriberón1,*, Ricardo Oliveros‐Ramos1, Erich Díaz1, Pierre Freón2 1 Instituto del Mar del Perú. 2 IRD, *[email protected] INTRODUCTION The catch per unit of effort (CPUE) of a fishing fleet has been proposed as an index of the abundance of the target stock, under the assumption that both are directly proportional, with proportionality determined by the catchability coefficient (q) (Baranov, 1918). However, in practice this proportionality is almost never true, because the catchability coefficient has temporal variability (Csirke, 1989), as result of the influence of the behavior of fish schools (Rose and Kulka, 1999) and of the fishing fleet. The objective of this work is to standardize the CPUE of the purse seine fleet of the Peruvian anchovy using Generalized Linear Models (GLM), and assess the performance of the standardized CPUE as an index of anchovy abundance.

This final model had R2 = 0.485, being the hold capacity (49.2%) and year (19.7%) the most influential variables in the model. The interaction with more influence in the variance of CPUE was year*month (10.9%). Model residuals were normally distributed.

Fig. 2. Correlation between the standardized CPUE of the purse seine fleet and the biomass estimated by a statistical model of cacth at age.

MATERIALS Y METHODOLOGY We use information from the “Programa Bitacoras de Pesca" of the Peruvian Marine Research Institute (IMARPE), corresponding to to the Northern ‐Central area of the Peruvian sea ( 4‐16°S) from 1996 to 2008. The CPUE used was tonnes of anchovy caught by hours of trip (t/hours). The CPUE was standardized using a Generalized Linear Model (GLM), using as explanatory variables the year, month, hold capacity (m3), latitude, spatial inertia and average distance to the coast. Additionally, interactions between variables were considered. In order to improve the distribution of the CPUE we apply a Box‐ Cox transformation, with parameter λ=0.315. We used a standard fleet with 400 m3 of hold capacity, latitude 8°S and spatial inertial of 5. A comparison between anchovy biomass (from a catch at age model) and CPUE (before and after the standardization) was carry on to assess the effect of the standardization over the performance of the CPUE as an abundance index for anchovy. RESULTS The distribution of the explained variable after Box‐Cox transformation is shown in Fig. 1. After a model selection, the best model include year, month, hold capacity, latitude, spatial inertia and the interaction between year*month, month*latitude and latitude*hold capacity as significant explanatory variables.

Fig. 1: Analysis of the distribution of the transformed CPUE (a) Histogram (b) Box‐plot; (c) Graphic Q‐Q plot.

Finally we observed that the standardized CPUE of the purse seine fleet had a significant correlation with the biomass (R= 0.74) (Fig. 2). DISCUSSION  We hoped that the variable year is the most influential in the model, because it is considered that this explains the pure annual variations of fish abundance (Hernández and Perrotta, 2006), but in the case of Peruvian anchovy are the features related to the ship contribute more in explaining the variability of the CPUE.

Ref: Baranov, I.E., (1918). On the question of the biological basis of fisheries. Izv.nauch.‐issl. ikhtiol.Inst., 1(1): 81–128 / Csirke, J. (1990). El uso de datos de esfuerzo y captura por unidad de esfuerzo en la investigación de recursos pelágicos en el instituto del mar del Perú (IMARPE). Doc.Int. IMARPE, 24 p. / Hernández, D. y Perrota, R., (2006). Influencia de las interacciones con el factor año en los índices anuales de abundancia obtenidos por modelos lineales generales utilizando datos de captura por unidad de esfuerzo. Rev. Invest. Desarr. Pesq. Nº 18: 57‐73. / Rose, G.A. , and Kulka D.W. (1999). Hyperaggregation of fish and fisheries: how catch‐per‐unit‐increased as the northern cod (Gadus morhua) declined. Can. J. Fish. Aquat. Sci. 56 (Suppl. 1): 118 – 127.

In a future work, we will include environmental variables (e.g. SST, depth) in order to explain a bigger fraction of the total variability of the CPUE, since this variables could explain important aspects of the behavior of Peruvian anchovy which are related with the catchability coefficient. Also, the ability and experience of the leader fishermen could have an effect in the CPUE and should be considered in a future work. The standardized CPUE reproduced better the fluctuation of the anchovy biomass (estimated from an independent source) than the original CPUE. CONCLUSIONS 1.

2. 3.

The best model explains a 50% of the CPUE variability, being the hold capacity the most important explanatory variable (49.2% of explained variance). The standardization of the CPUE improved its performance as anchovy abundance index. Environmental variables and fishermen experience should be tested as explanatory variables for the CPUE.