<<

"Not to be cited without prior reference to the authors" ICES CM 2007/Q:06

Chilean Assessment Surveys: A review.

Enzo Acuña1, R. Alarcón2, L. Cid3, L. Cubillos4 and R. León5.

Abstract

Three species are subject to in the lower shelf and upper slope (150 to 500 m) off Northern-Central : two galatheid squat ( johni and monodon) and a pandalid (Heterocarpus reedi). Actually, their fisheries are administered through a Quota system and therefore swept area surveys are conducted every year to provide tuning information for the VPA analysis and Quota calculation.

The different sampling designs that have been applied during the fourteen years of direct assessments, ranging from systematic transects to randomly stratified designs for the three resources, are critically analyzed. Methodological approaches in relation to tow duration, determination of tow beginning and end, geographic and bathymetric distribution of the three species, analysis of the by-catch caught with the three species are also addressed.

Keywords: Squat lobsters, Deepsea shrimp, Trawl research surveys, Chile.

1 Depto. de Biología Marina. Universidad Católica del Norte. Casilla 117- Coquimbo, Chile. Phone-Fax 56 51 209814, [email protected]. 2 Instituto de Investigación Pesquera Octava Región. Colón 2780. Talcahuano, Chile. Phone 56 41 2920410, Fax 56 41 2920411, [email protected]. 3 Depto. de Estadística. Universidad de Concepción. Concepción, Chile. Phone 56 41 2203177, Fax 56 41 2251529, [email protected]. 4 Depto. de Oceanografía. Universidad de Concepción. Concepción, Chile. Phone 56 41 2207233, Fax 56 41 2256571, [email protected]. 5 Edificio IdeaIncuba. Oficina Biocosta. Universidad de Concepción. Concepción, Chile. Phone 56 41 2797140, Fax 56 41 2207060, [email protected]. Introduction

The yellow squat , red and deepsea shrimp Heterocarpus reedi fisheries are important crustacean fisheries in northern- central Chilean lower shelf and upper slope, between 150 to 500 m depth. In general, the shallowest species is the red squat lobster which is found normally between 120 and 250 m, the yellow squat lobster is found between 180 and 300 m and the deepsea shrimp is found between 250 and 500 m depth and from 24 to 38°S (Fig. 1).

Since 1993, the Fisheries Research Fund (Fondo de Investigación Pesquera, FIP) has financed bottom trawl research surveys to determine the biomass and abundance of the three crustacean resources: the two squat lobsters and the deepsea shrimp found in northern-central Chilean waters. Different research teams have performed these studies and made some changes in the research protocol through time; therefore a review of these studies has been required by this financing agency.

The objective of determining the best sampling size is that taking into account survey costs and sampling variability, the unit size is found that produces the most precise density estimate given a fixed amount of survey resources or, if a certain level of precision is required, the size of sampling unit that minimizes the total cost of the survey (Pennington & Vølstad, 1991).

Bottom trawl surveys provide stock assessment modelers with -independent estimates of relative abundance based on catch per unit effort (CPUE), as well as age and size composition information. Precision of these CPUE estimates is in part affected by variability in the sampling efficiency of the trawl. To minimize this variability, survey scientists pay close attention to the maintenance of strict fishing procedures, such as gear deployment, retrieval, and towing speed (Weinberg et al., 2002).

One of the core issues in bottom trawl survey methodology and in other experimental trawling situations is standardizing and measuring the fishing effort. Whether survey results are used to prepare “area swept” estimates of absolute abundance, to compute indices of relative abundance, or to calculate other fishing efficiency statistics, controlling and quantifying tow duration, and tow distance are considered important. The significance of these factors is comparable to others, such as the construction and rigging of the gear, adherence to a specified towing speed, and other operational protocols (Stauffer, 2004). There are several examples of bottom trawl surveys programmes, like the MEDITS survey programme which are intended to produce basic information on benthic and demersal species in terms of population distribution as well as demographic structure, on the continental shelves and along the upper slopes at a global scale in the Mediterranean (Bertrand et al., 2002) which had as one of its main challenges the adoption of common standardized sampling protocols.

The sampling protocols for many groundfish trawl surveys define tow duration as the period between the time the trawl is determined to be on the bottom and in a stable fishing configuration and the time the trawl winches are engaged to retrieve the gear (usually at the end of some fixed, predetermined sampling period). In such cases, the towing distance is similarly defined as the distance transited by the vessel between the starting point and end point of the tow as previously defined (Wallace & West, 2006).

The swept area method requires quantitative information on the effective pathwidth of a trawl to estimate absolute densities of groundfish (Ramm & Xiao, 1995). Effort in trawl surveys is standardized by using a common gear and fixing haul duration and vessel speed. Such standardization should result in a fairly constant distance trawled and area or volume swept (Adlerstein & Ehrich, 2002).

Variability in trawl efficiency due to vessels and gears can be controlled by keeping these factors constant or adjusted for by applying corrections based on comparative fishing experiments when changes are made. Similarly, keeping surveys within similar time periods in each year can control seasonal variability. Conversely, diel variability can be more difficult to control, especially in the face of the logistical constraints of marine ecosystem sampling that often requires fishing to occur during both day and night (Benoît & Swain, 2003). Pennington & Vølstad (1991) pointed out several benefits obtained with decreasing tow duration, it saves survey time or more number of stations can be made with an increase in precision of abundance estimates, it reduces operating expenses, gear and equipment wear is a function of tow length, more fuel is consumed while dragging a trawl, there is less of a chance that an obstruction will cause a tow to be aborted or damage the gear, smaller catches which require less sorting time and allow more time for taking other biological measurements, gear saturation, the filling of the sampler with or debris before the tow is completed. Godø et al. (1990) analyzing data for cod, haddock and long rough dab had determined that short tows are at least as efficient as long tows in catching fish of any size and based on their results suggested that the efficiency of trawl surveys can be increased by reducing tow duration. Somerton et al. (2002) observed in the eastern Bering Sea (EBS) bottom trawl survey that, due to recent population increases of several fish species, 30 min tows often produced catches that were too large to be feasibly sorted in their entirety. These require subsampling procedures that are time-consuming and may result in increased sampling variability and bias, suggesting that one way to reduce the need for subsampling is to reduce tow duration.

Stocks of off West Greenland have been assessed using a research trawl survey since 1988 (Carlsson et al., 2000). The survey has used a design of randomly placed stations, stratified (on depth data where available, using small blocks elsewhere), with sampling effort proportional to stratum area. In some years, a two-stage adaptive sampling scheme was used to place more stations into strata with large first-stage variation in catches. The design of the survey was reviewed in 1998 (Carlsson et al., 2000, Folmer & Pennington, 2000). Four modifications to the survey were proposed during the first phase of the review: (1) increasing the number of stations sampled by shortening tow duration from 60 to 15 min, (2) reallocation of effort from areas of historically low density to areas of high abundance, (3) pooling strata to increase the sample size in the resulting (combined) strata and (4) abandonment of two-stage sampling. In 1998, tow duration was reduced from 1 h to 30 min at 25% of the stations, two- stage adaptive sampling was discontinued, and sampling effort was reallocated (Folmer & Pennington, 2000).

Relative abundance indices and their variances derived from trawl surveys were often estimated in the past from standard mean and variance calculations. This method is referred to as Normal in comparison with those in use for particular distributions. For a given species, hauls generally show an irregular distribution with many zero values and some very large catches (Caverivière, 1993). Presently, the Delta distribution system of analysis seems best suited to increase the efficiency in mean and variance calculations for trawl surveys (Pennington, 1983 and 1986). Delta distribution treats positive values separately on the assumption that they have a simple log- normal distribution, and then the zero values are included. A hyper-geometric function, which can easily be computed, is used. The efficiency of Delta distribution depends on the number of trawl tows, on the proportion of zero values and on the variability range for positive values (Smith, 1988).

Since the late 1980’s and in the 1990’s, geostatistics has been widely used for the estimation of abundance of demersal resources (Maynou, 1998). It has also been used by to study the spatial distribution of other galatheids like intermedia and M. sarsi on the Galician continental shelf in NW Spain (Freire et al., 1992) and for mapping, estimating biomass and optimizing sampling programs in the northern shrimp, Pandalus borealis (Simard et al., 1992). Over the past 10 years, fisheries scientists gradually adopted geostatistical tools when analysing fish stock survey data for estimating population abundance. First, the relation between model-based variance estimates and covariance structure enabled estimation of survey precision for non random survey designs. The possibility of using spatial covariance for optimising sampling strategy has been a second motive for using geostatistics. Kriging also offers the advantage of weighting data values, which is useful when sample points are clustered (Petitgas, 2001).

By-catch has long been recognized in fisheries, and in many instances it is unwanted and discarded. ‘‘Discards’’ is a term used for non-target species and undersized commercial animals that are rejected from the catch. has been estimated to produce 27 million t yr_1 discards world-wide (Alverson et al., 1994).

It is estimated that an unwanted by/catch of around 15% (by weight) was common in the Pandalus fisheries befor the introduction of the Nordmøre grid. Around 125 different species, notably including gadoids, redfish, turbot, sharks, Greenland halibut and capelin are discarded (Anon., 1998).

Otter trawl fishery for northern shrimp (Pandalus borealis) began in Atlantic Canada in 1965, where the small codend mesh (minimum 40-mm size is used) often resulted in high by-catches of many other species. In 1990 there was a total catch of 60,000 metric tones of shrimp with approximately 15% by-catch (by weight) (Anon., 1998).

Di Biasi (2004) used the PRIMER statistical software package to perform analyses of the community data. She used a cluster analysis, with the Bray-Curtis similarity index performed on fourth root transformed data. The resultant similarity matrix was used to perform non-metric multidimensional scaling (nMDS) to detect changes in the benthic assemblage due to the effect of fishing. The differences between seaward and landward controls before the experimental trawling were tested by the one-way ANOSIM. An a priori one-way ANOSIM-test was then performed to determine any significant differences between the fished and control treatments for each of the time intervals before and after the creation of the fishing disturbance. To establish which taxa contributed most to either the similarity or dissimilarity between groupings of data, the SIMPER routine was carried out.

This article reviews the characteristics of the surveys design and sampling strategy, tow duration, and in general the methodology used to assess the biomass and abundance of the two squat lobsters and the pandalid shrimp populations in Chilean waters and the improvements made through time in different aspects of the trawl surveys.

Material and Methods

Sampling strategies

Although several sampling strategies have been used, the randomization of the sampling selection needs to be conditioned by the spatial distribution of the species, which for these species has one of two characterizations. One in which the distribution of the species consists of clearly detectable fishing grounds (squat lobsters), and alternatively a continuous abundance band, in which it is difficult to distinguish fishing grounds (deepsea shrimp). In both cases, the fishing grounds and continuous abundance band were divided into sampling grids consisting of 1’x1’ cells, such that every one of these square cells was considered a sampling unit. Within each unit a sample was obtained by trawling as representative of the abundance of the entire cell.

Considering the restriction that the species are distributed in narrow bands parallel to the coast, running north to south along central Chile, historically (up to year 2003), the sampling strategies considered a systematic sampling with longitudinal transects. This had a twofold objective, to achieve a complete coverage of the area of the fishery, from 24 to 38°S with transects located every 10’ of latitude and secondly, to have an almost complete longitudinal coverage of the abundance bands (Figure 2.1.). The randomness of the selection was achieved by randomly selecting the location of the first transect. The transect consisted of a strip of sampling units with longitudinal orientation (east to west) to cover the entire abundance band width. Within this transect one of every two or every three units were trawled, and the abundance estimated for the tow was assumed to be the same for the entire unit.

A further refinement of the systematic sampling was the introduction of an adaptive strategy. This scheme was used due to the requirement of intensifying the sampling effort in the areas of higher abundance of the species. The adaptive strategy consisted of an increase of the number of transects in those areas where a predefined minimum capture for a given species was obtained. When such minimum was reached, two additional transects, located 5’ north and south of the main transect, were sampled (Figure 2.2.). The same procedure was used in all cases where the capture was higher than the minimum. Although this is a well documented procedure, the two main drawbacks were that this type of strategy is designed for clustered populations (Francis 1984, Thompson 1988, 1990, 1991a, 1991b, 1992, Thompson et al. 1992), and secondly, that it is not possible to previously define the sample size, which is a major issue in a research where the budget does not allow but a certain amount of tows. This dynamic allocation of stations also had some logistic inefficiencies of extra steaming and back-tracking to add stations in areas already sampled, as was also observed by Carlsson et al. (2000) in the Pandalus borealis fishery off Greenland, therefore this allocation was abandoned and only a predetermined design was used.

Since in some cases species are not distributed in clusters, but in a more uniform way along their abundance bands, systematic sampling was not considered to be an appropriate strategy, then simple random sampling was used along the entire band of abundance (Figure 2.3.). This strategy has been documented as more effective than systematic, stratified random and adaptive sampling for the cases of uniform distribution (Cochran, 1977, Thompson, 1992).

In contrast with previous randomization strategies, stratified random sampling has been successfully used in the last four years for the squat lobster species, which have clearly delimited fishing grounds. In fact, in most cases there exists a requirement of intensifying the sampling effort precisely in those fishing grounds. For that purpose, a stratified randomization was defined, using the historic fishing grounds determined with previous data as strata, being each fishing ground defined as a different stratum. The number of samples was allocated in each stratum proportional to its size and using simple random sampling within the stratum.

As an alternative stratification method, the abundance gradient of the species was used to define the strata. In such cases, the allocation of number of samples was determined by the species density instead of the size of the strata, as shown in Figure (3). In this case, the abundance gradient has a latitudinal orientation.

Tow duration

To implement some of the benefits listed by Pennington and Vølstad (1991) and Godø et al. (1990) in Chilean crustacean research trawl surveys, shorter tows were implemented since 2003, replacing the 30 min tows previously used, by 15 min tows. The trawl bottom contact, which is considered the beginning of ”the effective tow duration” or towing time, was determined by acoustic instrumentation (NETMIND tilt sensor) (Figure 4).

Sampling gear and onboard procedures

The standard device is a bottom trawl, including all the material and its rigging from the doors to the codend of the net. Its codend mesh size is 40 - 50 mm (stretched mesh). The chosen fishing gear used by the vessels is the same bottom trawl used commercially and has remained with almost no modifications over the history of the surveys (Figure 5).

Along the coast two different areas are sampled, a strip of 5 nautical miles (nm) which is reserved by law for the artisanal fishermen, where an artisanal boat i.e. less than 18 m total length must be used, and the area outside the previous one where industrial vessels can be used. Chartered fishing vessels are used, and as much as possible, the same vessels are used every year in each area, what has happened during the last four years when the same industrial vessel has been used for the surveys.

The depth of the sampling locations varied from 50 to 450 m depth allowing, encompassing the whole depth range of the two squat lobsters in the area of study. In the case of the deep sea shrimp, the set depth of the sampling locations varied from 100 to 600 m depth allowing, encompassing the whole depth range of this species.

The total weights of the target species and their by-catch are obtained on board, but the samples from the target species and their by-catch for biological purposes are only taken and preserved on board and all the biological analyses are carried out on land in the laboratories. Observations on the target species are length frequency distribution from the sample taken of each tow and individual weight and sex (including sexual maturity stage) of a subsample. Fecundity is also normally determined. Spatial distribution of target species

In the geostatistical literature, an autocorrelated variable is termed a regionalized variable (Maynou, 1998). The example of regionalized variable which is used in our bottom trawl surveys is the density of P. monodon or C. johni over the fishing grounds off northern-central Chile.

The fishing grounds limits were determined applying the “Transitive geostatistical method” and the analysis was performed separately and independently for both squat lobster species. The scientific support of the use of this scope is based in the fact that the sampling design (tows) surpass the limits of the species distribution and that all sample values are considered, including zeros. We also assumed that the population densities systematically decrease towards the limits of the species distribution.

Biomass and abundance estimates

Several methods have been used to determine the biomass and abundance of the two squat lobsters or deepsea shrimp: Δ distribution, geostatistical analysis,

The covariance function describing the spatial autocorrelation is called structure function (Maynou, 1998), is this case the variogram. The procedure to compute the geostatistical experimental variograms used follows Maynou (1998) in particular the procedure described in the text and shown in his Figure 2.

By-catches studies

Each species was described, bathymetrically (every 50 to 100 m depending upon the tow distribution) and latitudinally, using its relative abundance (Capture per Unit of Area, CPUA), frequency of occurrence in the total positive tows and relative importance (%) with respect to the total by-catch and to the target species.

Spatially the specific assemblages will be identified by means of the Cluster and MDS Analysis (Clifford & Stephenson, 1975), using the software PRIMER (Plymouth Marine Laboratory; Clarke & Warwick, 1994). This analysis consists in comparing tow pairs, using the cpue of each species, using the Bray-Curtis Similarity Index (IBC, Bray & Curtis, 1957). With these values a similarity matrix is built which will be used for the Cluster and MDS Analysis. To balance the rare species values with the common species values the original data (cpue) is transformed with the fourth root according to Clarke & Warwick (1994).

Once the different species assemblages are identified through the Classification Analysis they are compared with ANOSIM (Analysis of Similarities) (Clarke & Green, 1988) procedures, implemented in the software PRIMER.

Results

Only some selected results for the two squat lobster species as an example are shown, since the most dramatic change due greatly to the change in sampling design has been found in these two species and the main purpose of this paper is to review some of the aspects involved in the changes implemented in the bottom trawl crustacean surveys.

The impact of the stratified random sampling design in comparison with the adaptive design previously used in terms of the number of species positive samples, i.e. samples where the red or yellow squat lobsters were captured are shown in Table 1. In both species the number of species positive samples obtained were higher, but in 2005 for P. monodon.

Besides the number of positives samples i.e. samples where one or the two species of squat lobsters were captured, that were obtained with the stratified random sampling design was always higher than 67.9%, but unfortunately was not available from previous studies for comparison purposes (Table 1).

The fishing grounds limits for both squat lobster species in part of the study area are shown in Figures 6 and 7, where three different ways of representing these areas are presented.

The by-catched species with higher relative importance, with respect to the total by-catch CPUA and the total CPUA of the target species (deepsea shrimp and squat lobsterss), are only two or three species (Fig. 8). In the years analyzed the species with higher relative importance were the common hake (Merluccius gayi) and the bigeye flounder (Hippoglossina macrops) (Fig. 8). Other species that have lower relative importance become relevant for having a higher frequency of occurrence than the rest, like the Mursia gaudichaudii, in year 2004 and Cancer porteri, in years 2005 and 2006.

The species that determine the group internal similarity repeat themselves every year, being the two most important the common hake Merluccius gayi and the bigeye flounder Hippoglossina macrops (Table 2 A-C). Other species that are present in this analysis in some years are the jumbo squid Dosidicus gigas, the grenadiers Nezumia pulchella and Caelorinchus aconcagua, the spider Libidoclea granaria and the crab Mursia gaudichaudii (Table 2 A-C).

Discussion

The methodological changes seen in the bottom trawl research surveys to assess the biomass and abundance of lower shelf and upper slope crustacean in the northern-central Chilean waters, especially in terms of the sampling design are very similar to those reported by Carlsson et al. (2000) for the research trawl survey of the stocks of Pandalus borealis off West Greenland.

The issues addressed, up to this point, in the Chilean experience are the sampling design, the number of stations (tows), tow duration, reallocation of effort from areas of historically low density to areas of high abundance, several of these also suggested by Folmer and Pennington (2000) in their analysis of the research trawl survey of the stocks of Pandalus borealis off West Greenland.

Issues related with the sampling gear have not been addressed yet and only the commercial trawl has been used, however, the same trawl as well as the same fishing vessel has been used during the last four years. Actually there is an ongoing study analyzing the commercial trawl characteristics which will probably end with a new design for the trawl, which should then be studied with respect to the several aspects related with the sampling gear involved as analyzed by Wallace & West (2006), Ramm & Xiao (1995) and Adlerstein & Ehrich (2002) among others.

An ongoing analysis of the swept area studies performed in Chilean waters since 1993 has the aim to propose a Manual, where all the characteristics of the bottom trawl research studies should be considered with respect to sampling design, the information collected, and the management of the data as far as the data storage (Data Bank) and production of common standardized analyses of the data. There are several examples of bottom trawl surveys programmes that have produced this kind of Manuals, like the MEDITS survey programme (Bertrand et al., 2002, MEDITS, 2007) had as one of its main challenges the adoption of common standardized protocols, the IBTS Group (ICES, 1999) and the Manual of the Baltic International Trawl Surveys (BITS) (ICES, 2007).

References

Acuña, E., G. Conan, L. Cid, R. Alarcón & L. Cubillos. 2004. Evaluación directa de colorado entre la III y IV Regiones, año 2003. Informes Técnicos FIP. FIP/IT Nº 2003-03, 141 pp.

Acuña, E., R. Alarcón, H. Arancibia, L. Cid, A. Cortés, L. Cubillos, R. León y S. Neira. 2005. Evaluación directa de langostino colorado y langostino amarillo entre la II y VIII regiones, año 2004. Informes Técnicos FIP, FIP/IT Nº 2004-11, 398 pp.

Acuña, E., R. Alarcón, H. Arancibia, L. Cid, L. Cubillos y A. Cortés. 2006. Evaluación directa de langostino colorado y langostino amarillo entre la II y VIII Regiones, año 2005. Informes Técnicos FIP. FIP/IT Nº 2005-09, 348 pp.

Acuña, E., R. Alarcón, H. Arancibia, L. Cid, A. Cortés, Luis Cubillos y Rafael León. 2007. Evaluación directa de langostino colorado y langostino amarillo entre la II y VIII Regiones, año 2006. Informes Técnicos FIP. FIP/IT Nº 2006-04, 375 pp.

Adlerstein, S. and S. Ehrich. 2002. Effect of deviations from target speed and of time of day on catch rates of some abundant species under North Sea International Bottom Trawl Survey protocol conditions ICES Journal of Marine Science, 59: 594–603.

Alverson, D. L., M. H. Freeberg, S. K. Murawski and J. G. Pope. 1994. A global assessment of fisheries by-catch and discards. Food and Agriculture Organization, Rome. 233 pp.

Anon. 1998. Report of the Study Group on Grid (Grate) Sorting Systems in Trawls, Beam Trawls and Seine Nets. . ICES CM 1998/B: 2, 63 p.

Bahamonde, R., B. Leiva, C. Canales, M.A. Barbieri, J. Cortes, J.C. Quiroz, P. Arana, A. Guerrero, M. Ahumada, T. Melo, D. Queirolo, C. Hurtado, P. Gálvez y E. Molina. 2004. Evaluación directa de langostino colorado y langostino amarillo entre la II y VIII Regiones, año 2002. Informes Técnicos FIP. FIP/IT Nº 2003-31, 325 pp.

Benôit, H.P. and D.P. Swain. 2003. Accounting for length- and depth-dependent diel variation in catchability of fish and invertebrates in an annual bottom-trawl survey. ICES Journal of Marine Science, 60: 1298–1317.

Bertrand, J.A., L. Gil De Sola, C. Papaconstantinou, G. Relini and A. Souplet. 2002. The general specifications of the MEDITS surveys. Sci. Mar., 66 (Suppl. 2): 9-17.

Bray, J. R. & J. T. Curtis, 1957. An ordination of the upland forest communities of Southern Wisconsin. Ecol Monogr., 27: 325-349.

Canales, C. M.A. Barbieri, R. Bahamonde, B. Leiva, P. Arana, S. Palma and T. Melo. 2002. Evaluación directa de langostino colorado y langostino amarillo entre la II y VIII Regiones, año 2001. Informes Técnicos FIP. FIP/IT Nº 2001-06, 269 pp.

Canales, C. M.A. Barbieri, R. Bahamonde, B. Leiva, P. Arana, A. Guerrero, M. Ahumada, T. Melo, D. Queirolo, C. Hurtado, P. Gálvez, S. Palma, E. Molina y N. Silva. 2003. Evaluación directa de langostino colorado y langostino amarillo entre la II y VIII Regiones, año 2002. Informes Técnicos FIP. FIP/IT Nº 2002-06, 410 pp.

Carlsson, D. M., Folmer, O., Kanneworff, P., Kingsley, M. C. S., and Pennington, M. 2000. Improving the West Greenland trawl survey for Pandalus borealis. Journal of Northwest Atlantic Fishery Science, 27: 151-160.

Caverivière, A. 1993. Some methodological considerations on Delta distribution, stratification and tow duration, for trawl surveys carried out in West Africa. Fisheries Research, 16: 223-237.

Clarke K.R. & R.H. Green, 1988. Statistical design and analysis for a biological effects study. Mar Ecol. Prog. Ser., 92: 205-219.

Clarke, K. R. & R. M. Warwick, 1994. Changes in marine communities: an approach to statistical analysis and interpretation. Plymouth: Plymouth Marine Laboratory, 144 pp.

Cochran, W. G. 1977. Sampling Techniques. John Wiley & Sons, New York.

De Biasi, A. M. 2004. Impact of experimental trawling on the benthic assemblage along the Tuscany coast (north Tyrrhenian Sea, Italy). ICES Journal of Marine Science, 61: 1260-1266.

Escuela de Ciencias del Mar. 2000. Evaluación directa de camarón nailon, langostino amarillo y langostino colorado, año 2000. Informes Técnicos FIP, FIP/IT Nº 2000-05, 315 p.

Folmer, O., and Pennington, M. 2000. A statistical evaluation of the design and precision of the shrimp trawl survey off West Greenland. Fisheries Research, 49(2): 165-178.

Francis, R.I.C.C., 1984. An adaptive strategy for stratified random trawl surveys. Journal of Marine and Freshwater Research. 18:59-71.

Freire, J., E. González Gurriarán and I. Olaso. 1992. Spatial distribution of Munida intermedia and M. sarsi (Crustacea: ) on the Galician continental shelf (NW Spain): Application of geostatistical analysis. Est. Coast. Shelf Sci., 35: 637-648.

Godø, O.R., Pennington, M. and J.H. Vølstad. 1990. Effect of tow duration on length composition of trawl catches. Fish. Res. 9: 165-179.

ICES, 1999. Manual for the International Bottom Trawl Survey - Revision VI. ICES CM 1999/ D: 2, Addendum 2.

ICES, 2007. Manual of the Baltic International Trawl Surveys (BITS). Rostock, Germany, 71p.

Maynou, F. 1998. The application of geostatistics in mapping and assessment of demersal resources. (L.) in the northwestern Mediterranean: a case study. Sci. Mar. 62 (Supl. 1): 117-133.

MEDITS, 2007. International bottom trawl survey in the Mediterranean (MEDITS). Instruction manual. Version 5, 62 p.

Pennington, M. 1983. Efficient estimators of abundance, for fish and surveys. Biometrics. 39: 281-286.

Pennington, M., 1986. Some statistical techniques for estimating abundance indices from trawl surveys. Fish. Bull., 84(3): 519-525.

Pennington, M., and J.H. Vølstad. 1991. Optimum size of sampling unit for estimating the density of marine populations. Biometrics, 47: 717–723.

Petitgas, P. 2001. Geostatistics in fisheries survey design and stock assessment: models, variances and applications. Fish and Fisheries 2: 231–249.

Ramm, D.C. and Y. Xiao. 1995. Herding in groundfish and effective pathwidth of trawls. Fisheries Research 24: 243-259.

Simard Y, Legendre P, Lavoie G, Marcotte D. 1992. Mapping, estimating biomass and optimizing sampling programs for spatially autocorrelated data: case study of the northern shrimp (Pandalus borealis). Canadian Journal of Fisheries and Aquatic Sciences 49: 32-45.

Smith, S.J., 1988. Evaluating the efficiency of the delta-distribution mean estimator. Biometrics 44: 485-493. Somerton, D.A., R.S. Otto and S.E. Syrjala. 2002. Can changes in tow duration on bottom trawl surveys lead to changes in CPUE and mean size? Fisheries Research 55: 63–70.

Stauffer, G. (Compiler), 2004. NOAA protocols for groundfish bottom trawl surveys of the nation’s fishery resources. U.S. Dept. Commerce, NOAA Tech. Memo. NMFS-F/SPO-65, pp. 81–102.

Thompson, S.K., 1990. Adaptive cluster sampling. Journal of the American Statistical Association 85: 1050-1059.

Thompson, S.K., 1991a. Adaptive cluster sampling. Designs with primary and secondary units. Biometrics, 47: 1103-1115.

Thompson, S.K., 1991b. Stratified adaptive cluster sampling. Biometrika, 78: 389-397.

Thompson, S.K., 1992 Sampling. Wiley and Sons Inc. New York.

Thompson, S.K., F. Ramsey and G.A.F. Seber. 1992. An adaptive procedure for sampling populations. Biometrics 48: 1195-1199.

Wallace, J.R. and C.W. West. 2006. Measurements of distance fished during the trawl retrieval period. Fisheries Research 77: 285–292.

Weinberg, K. L., Somerton, D. A., and Munro, P. T. 2002. The effect of trawl speed on the footrope capture efficiency of a survey trawl. Fisheries Research, 58: 303–313.

Table 1. Number of positive samples, i.e. samples where one or the two species of squat lobsters were captured and number of species positive samples, i.e. samples where the red (P. monodon) or yellow (C. johni) squat lobsters were captured, obtained with different sampling designs used in trawl research surveys off Chile. *= not available.

Sampling design Tows Project number and citation Total Positive P. monodon FIP 2000-05 Esc. Cs. del Mar (2000) Adaptive / transects 792 * 186 (23.5%) FIP 2001-06 Canales et al. (2002) Adaptive / transects 682 * 228 (33.4%) FIP 2002-06 Canales et al. (2003) Adaptive / transects 1.168 * 158 (13.5%) FIP 2003-31 Bahamonde et al. (2004) Adaptive / transects 719 * 127 (17.7%) FIP 2003-03 Acuña et al. (2004) Randomly Stratified 271 188 (69.4%) 115 (42.4%) FIP 2004-11 Acuña et al. (2005) Randomly Stratified 876 595 (67.9%) 266 (44.7%) FIP 2005-09 Acuña et al. (2006) Randomly Stratified 807 594 (73.6%) 196 (24.3%) FIP 2006-04 Acuña et al. (2007) Randomly Stratified 847 650 (77.8%) 288 (34.0%)

Sampling design Tows Project number and citation Total Positive C. johni FIP 2000-05 Esc. Cs. del Mar (2000) Adaptive / transects 792 * 298 (37.6%) FIP 2001-06 Canales et al. (2002) Adaptive / transects 682 * 266 (39.0%) FIP 2002-06 Canales et al. (2003) Adaptive / transects 1.168 * 330 (28.3%) FIP 2003-31 Bahamonde et al. (2004) Adaptive / transects 719 * 276 (38.4%) FIP 2004-11 Acuña et al. (2005) Randomly Stratified 876 595 (67.9%) 397 (45.3%) FIP 2005-09 Acuña et al. (2006) Randomly Stratified 807 594 (73.6%) 327 (40.5%) FIP 2006-04 Acuña et al. (2007) Randomly Stratified 847 650 (77.8%) 368 (43.5%)

Table 2. Species that determine the group of tows internal similarity, for the years (A) 2004 (B) 2005 and (C) 2006 (FIP 2004-11, 2005-09 y 2006-04, respectively). Merlgay = Merluccius gayi; Hippmac = Hippoglossina macrops; Dossigig = Dosidicus gigas; Nezulpul= Nezumia pulchella; Libigra= Libidoclea granaria; Caelaco= Caelorinchus aconcagua; Mursgau= Mursia gaudichaudii.

Figure 1. Crustacean bottom trawl fisheries in northern-central Chile lower shelf and upper slope (22 to 38°S). Red squat lobster (Pleuroncodes monodon), Yellow squat lobster (Cervimunida johni) and deepsea shrimp (Heterocarpus reedi).

Figure 2. Sampling designs used in the bottom trawl surveys to assess crustacean resources in Chile. 1. Systematic transects. 2. Adaptive transects and 3. Stratified random sampling. CPUE (kg/km) H. reedi 0 100 200 300

-36 -34 -32 -30 -28 -26 -24 Latitud S

Figure 3. Latitudinal variation of H. reedi CPUE (kg/ km) analysis used to define sampling stratification in relation to local abundance.

Figure 4. Screen view of measurements of wing spread and beginning of bottom contact of the trawl by acoustic instrumentation (NETMIND tilt sensor). (10,0) (12,0)

(5,0)

(200) (4,0)

(150)

11 – 15 PL φ 8”

13.00 Tonina 5 /8” 6.00 Tonina 5 / 8” 13.00 Tonina 5 / 8” 8 - 10 Tonina 5 / 8” ooooo 11.00 Tonina 3 / 4” Cadena 3.00 Tonina 5 / 8” FE 330–360 Kg

Figure 5. Trawl used in the crustacean research assessments in Chilean waters. Longitud (ºW) Longitud (ºW) 72.0 71.5 72.0 71.5 29.0 29.0

40 30 30 28 25 26 20 24 15 22 10 20 9 18 29.5 29.5 8 16 7 14 6 12 5 10 4 8 Latitud (ºS) Latitud 3 6 (ºS) Latitud 2 4 1 2 30.0 30.0 0 0

30.5 30.5

Figure 6. Fishing grounds (densities, ton/km2) of squat lobsters in the crustacean research assessments in northern Chilean waters.

Figure 7. Fishing grounds (densities, ton/km2) of squat lobsters in the crustacean research assessments in northern Chilean waters, in 3D and contours.

Figure 8. By-catched species Relative Importance in years 2004 to 2006 (FIP 2004-11, 2005-09 y 2006-04, respectively). IRFA= Relative Importance with respect to all species by- catched, IRRO= Relative Importance with respect to target species.