Quantifying light-fishing for Dosidicus gigas in the eastern Pacific using satellite remote sensing

Transcription

1 Remote Sensing of Environment 91 (2004) Quantifying light-fishing for Dosidicus gigas in the eastern Pacific using satellite remote sensing C.M. Waluda a, *, C. Yamashiro b, C.D. Elvidge c, V.R. Hobson d, P.G. Rodhouse a a British Antarctic Survey, National Environmental Research Council, High Cross, Madingley Road, Cambridge CB3 0ET, UK b Instituto del Mar del Perú, Esquina Gamarra y General Valle Chucuito, Apartado 22 Callao, Peru c Office of the Director, NOAA National Geophysical Data Center, 325 Broadway, Boulder, CO 80303, USA d Cooperative Institute on Research in the Environmental Sciences, University of Colorado, Boulder, CO 80303, USA Received 20 October 2003; received in revised form 11 February 2004; accepted 18 February 2004 Abstract The distribution and abundance of the fleet targeting Jumbo flying squid (Dosidicus gigas) in the Eastern Pacific is examined during the 1999 fishery season. The commercial fishery consists of a multinational jigging fleet, which fish at night using powerful lights to attract squid. The emission of light from these vessels can be observed using satellite-derived imagery obtained by the United States Defence Meteorological Satellite Program-Operational Linescan System (DMSP-OLS). In order to quantify fishing effort using lights, data on the distribution and abundance of vessels were obtained via satellite tracking using the ARGOS system. The distribution of the fishery as derived from light signatures was found to closely resemble that derived from ship location data. By using ARGOS data to calibrate DMSP-OLS images, we are able to estimate fishing effort in terms of the area illuminated by the fishing fleet. Light signatures derived from DMSP-OLS were successfully used to quantify fishing effort, estimating the number of vessels fishing to within F2 in 85 out of 103 satellite images (83%). High seas fishing was also quantified, with light signatures corresponding to a single fishing vessel observed in 11 out of 103 satellite passes during the fishery season (July December 1999). This study examines how much light (in terms of area) is emitted by a single squid fishing vessel, and may prove to be a valuable tool in assessing and policing fisheries using satellite remote sensing. D 2004 Elsevier Inc. All rights reserved. Keywords: Squid fishery; Dosidicus gigas; DMSP-OLS; ARGOS; Remote sensing; GIS 1. Introduction Dosidicus gigas, the Jumbo flying squid, supports a major fishery in the Eastern Pacific, off the coasts of North and South America (Nigmatullin et al., 2001). This species has a semi-oceanic pelagic habitat, and occurs at depths of up to 1200 m (Nigmatullin et al., 2001). D. gigas is distributed from 37 40jN to 45 47jS, between California and Southern Chile, extending westwards from the coast to a maximum of jW at the equator (Nesis, 1983; Nigmatullin et al., 2001). Since 1991, this species has been targeted in the East Central and South East Pacific (FAO areas 77 and 87) by a multinational jigging fleet (Ichii et al., 2002; Yamashiro et al., 1998), with a resulting increase in catches from 14,893 tonnes in 1990, to 223,784 tonnes in 2001 (FAO, 2000). * Corresponding author. Tel.: ; fax: address: clwa@bas.ac.uk (C.M. Waluda). Squid jigging vessels operate at night using powerful lights (up to 300 kw per vessel) to attract squid (Rodhouse et al., 2001). Using remotely sensed data from the United States Defence Meteorological Satellite Program-Operational Linescan System (DMSP-OLS) it is possible to observe the distribution of fishing fleets around the globe by means of the light emitted during fishing operations (Rodhouse et al., 2001). The distribution of lights can be used to observe the distribution and abundance of squid jiggers, and by inference, the distribution of exploited squid stocks, and the location of favourable fishing grounds (Cho et al., 1999; Waluda et al., 2002). Recent work using DMSP-OLS data suggested the distribution of the fleet operating off the Pacific coast of Central and South America was not influenced by bathymetry, and that the fishery targeted a distinct fishing area from year to year (during 1994, 1996 and 1997) within high seas and Exclusive Economic Zone (EEZ) waters (Waluda & Rodhouse, in press). In this study, we examine the fishery for D. gigas off the coast of Peru and in the adjacent high seas during the /$ - see front matter D 2004 Elsevier Inc. All rights reserved. doi: /j.rse

2 130 C.M. Waluda et al. / Remote Sensing of Environment 91 (2004) location of fishing lights derived using DMSP-OLS. The location of the fleet as derived from these two data sources are compared on an annual and monthly basis, and the area illuminated per vessel calculated. A regression analysis comparing illuminated area with number of vessels fishing was used to quantify the fleet using lights. Using this calibration, the incidence of fishing on the high seas off the coast of Peru was also examined and quantified. 2. Materials and methods 2.1. Data sources Fig. 1. Map of the Eastern Pacific. Boxed region indicates area from which Defense Meteorological Satellite Program-Operational Linescan System (DMSP-OLS) and ARGOS data describing the distribution of the D. gigas light-fishing fleet were derived. Dashed line indicates 200 nautical mile Exclusive Economic Zone (EEZ). fishing season using satellite remote sensing and Geographic Information Systems (GIS) techniques. Data on the distribution of individual vessels were obtained via satellite tracking using the ARGOS system, and compared with the DMSP-OLS data A series of 103 single-pass visible band images were obtained from the Defence Meteorological Satellite Program-Operational Linescan System (DMSP-OLS) for the region shown in Fig. 1 (0 20jS, 75 95jW). Images were selected for the period July to December 1999, corresponding to the period when the D. gigas fishery was operational. All satellite passes were made between the hours of 00:01 and 01:32 local time. Images were georeferenced using algorithms developed by the National Geophysical Data Center (NGDC), Boulder, CO (Elvidge et al., 1999) and incorporated into a Geographic Information System (GIS; Arc/Info version 8.1 ESRI, Environmental Systems Research Institute) where they were converted to a mercator equal-area projection with a resolution of approximately 1 km. The DMSP-OLS visible band has 6-bit quantization, producing digital number (DN) values ranging from 0 (no radiance) to 63 (saturated radiance) (Elvidge et al., 1999). For each image, pixels with a DN value of z30 (i.e. the brightest 50% of lit pixels, assuming an even split Fig. 2. Distribution of the D. gigas light-fishing fleet (July December 1999) using data from DMSP-OLS (left panel), and ARGOS (right panel). Solid line indicates 200 and 1000 m bathymetric contours.

3 C.M. Waluda et al. / Remote Sensing of Environment 91 (2004) ARGOS data Satellite-tracked data from the fleet fishing in Peruvian waters (at a resolution of approximately 1 km) were obtained using Platform Transmitter Terminals (PTTs) attached to fishing vessels and received via ARGOS instruments flown aboard National Oceanic and Atmospheric Administration (NOAA) Polar Orbiting Environmental Satellites (POES). All vessels fishing for D. gigas off the coast of Peru during 1999 were fitted with a PTT. Data on the location of each vessel (latitude, longitude and time) were obtained via a regional Collect Location System (CLS) by the Instituto del Mar del Perú (IMARPE), Callao, Peru for each fishing vessel, for the period 3 July 31 December 1999 (182 nights). To correspond with the timing of DMSP- OLS data, only those records received during the same time period (00:01 01:32 h inclusive) were used in the analysis. Between 46 and 134 records were obtained per month, corresponding to a total of between 1 and 14 vessels recorded per night Analyses Fleet distribution during 1999 The distribution of the squid fleet fishing off the coast of Peru was examined during July to December No fishing took place during January to June of this year. The location of the fishing region as described by (a) illuminated area derived from DMSP-OLS data and (b) ARGOS data was examined on both an annual and monthly basis in order to examine patterns of distribution over the fishing season and make comparisons between the two data sources Calibration and estimation of vessel numbers in Peruvian waters Using data on the number of vessels fishing per night (derived from ARGOS) and data on the number of illuminated pixels observed off the coast of Peru (derived from DMSP-OLS), a regression equation relating the number of pixels to the number of vessels was derived, in order to compare area illuminated with size of fleet. Using a GIS, the area illuminated was calculated for each image and divided Fig. 3. Monthly distribution of the D. gigas fleet off the coast of Peru, July December DMSP-OLS data shown in yellow; ARGOS positions shown in red. between lights from vessels and lights reflected from the ocean surface) were extracted and defined as being representative of fishing lights. Images were excluded from the analysis if there was substantial contamination by solar or lunar glare, or if heavy cloud obscured the fishing region. Fig. 4. The number of vessels fishing in Peruvian waters versus the number of illuminated pixels in corresponding DMSP-OLS images. ( ) Regression line (R 2 =0.72), error bars indicate 95% confidence intervals (regression equation: N vessels =2.467+(0.03N pixels ); P<0.001; n=103).

4 132 C.M. Waluda et al. / Remote Sensing of Environment 91 (2004) by the number of vessels fishing in order to give an estimate of the amount of light emitted per vessel, and quantify fishing effort using lights from DMSP-OLS. The occurrence of fishing lights on high seas waters were also recorded, and an estimation of the number of vessels fishing in this region was made, based on the calibration described above. which is consistent with the signature for a single vessel (Fig. 4). Lights on the high seas often occurred close to the 200 nautical miles EEZ and of particular note, a single vessel was observed in several locations close to the EEZ between 8 and 15 October Fishing on the high seas was not recorded using data from ARGOS. 3. Results 3.1. Fleet distribution during 1999 During 1999, the squid jigging fleet were observed to be operating between 3 and 7jS, and from the coast to around 83jW (Fig. 2). The fishery was located offshore of the 1000 m bathymetric contour, with fishing taking place between 20 and 120 nautical miles from the shore. Both data sources revealed a similar fishing area (Fig. 2). Throughout the season, the fishery was centred at around 5jS, with occasional fishing occurring to the south of this region. The fleet appeared to move south as the season progressed (Fig. 3) Calibration and estimation of vessel numbers in Peruvian waters Using the calibration (correlation coefficient r =0.86) given in Fig. 4 to examine fishing intensity as indicated by the area illuminated, it was calculated that number of vessels fishing in Peruvian waters varied between 3 and 17 (mean=5f3; Fig. 5). The number of vessels fishing as derived from ARGOS data varied from 1 to 6 between July and November, increasing up to 14 vessels in December (mean=6f3; Fig. 5). In Peruvian waters, the number of vessels estimated using DMSP-OLS was correct on 24% of occasions (25 images), and to within F2 vessels on 83% of occasions (85 images) (Fig. 5). The area illuminated per vessel fishing off the coast of Peru was calculated to be between 1 and 7 km 2 (mean=3.9f1 km 2 ). Fishing outside the Peruvian EEZ was observed on 11 occasions using DMSP-OLS data. In each case, the area illuminated was calculated to be between 2.5 and 5 km 2 (mean 4F1 km 2 ), Fig. 5. Number of vessels fishing off the coast of Peru, July December Actual number of vessels derived from ARGOS data (black circles); estimated number of vessels derived from DMSP-OLS images (grey triangles) using the regression equation shown in Fig Discussion The location of the fishing fleet was successfully observed using satellite remote sensing data from the Defence Meteorological Satellite Program-Operational Linescan System (DMSP-OLS) and satellite tracking using ARGOS. The use of DMSP-OLS data appears to give a valid indication of the distribution and abundance of the fleet fishing off the coast of Peru. ARGOS data are available more frequently and cover the whole of the fishing season, but DMSP-OLS data seem to be a good proxy for examining the spatial distribution of the fleet. Both sets of data are automatically received such that it is impossible to misreport the location of a vessel, making satellite monitoring a powerful tool in the policing of fisheries, and the estimation of fishing effort on the high seas. The distribution of the fleet in 1999 was similar to that observed in previous years as indicated by both fishery observer data (Taipe et al., 2001; Yamashiro et al., 1998) and DMSP-OLS data (Waluda & Rodhouse, in press), and it appears that the squid population inhabits a fairly consistent region from year to year, at least off the coast of Peru (Taipe et al., 2001; Waluda & Rodhouse, in press; Yamashiro et al., 1998). The aggregation of squid in this region is likely to be related to areas of high productivity linked to the upwelling system off Peru (Anderson & Rodhouse, 2001; Longhurst, 1998; Thomas et al., 1994), and the associated productivity of prey species, such as myctophids and other mesopelagic fishes, which constitute a large part of the diet of D. gigas (Nigmatullin et al., 2001). Throughout the year, the fleet was centred at around 5jS. Some fishing in the south of the region (to around 8jS) was observed, but this was limited and occurred only during July, August and October. There is some suggestion that the fleet moved gradually southwards as the season progressed, which may relate to the smallscale migration of squid. The use of illuminated area as a measure of fleet size appears to be a good estimator of vessel number. Some degree of variability was observed in the light signature emitted by a single vessel, which may potentially be attributed to factors such as atmospheric variability, angle of satellite, high cloud cover and timing of satellite pass. Also, the light signature from several vessels fishing close together may give an underestimation of the area illuminated per vessel due to an overlap of light signatures. The area illuminated by a single vessel is much greater than the size of the vessel itself, and although we have accounted for light reflected from the ocean surface, this is difficult to quantify

5 C.M. Waluda et al. / Remote Sensing of Environment 91 (2004) accurately using the current techniques. Further work should therefore examine alternative methods of radiance calibration, and compare the radiance of individual vessels with the output from DMSP-OLS imagery. The calibration comparing vessel numbers with illuminated area yielded similar results to a previous analysis examining fishing in the South Atlantic (Waluda et al., 2002), suggesting that this technique is transferable to other light-fisheries for squid. The number of vessels fishing per night in Falkland Islands waters was found to be much higher than in Peruvian waters (a mean of 75 vessels compared to 6 in the present study), but the same vessels are likely to be involved in each fishery, particularly as the Southeast Pacific season directly follows the end of the season in the Southwest Atlantic, which usually closes in June. Light signatures received via DMSP-OLS should thus be similar for both fisheries. High seas fishing in the Southeast Pacific was occasionally observed in DMSP-OLS imagery, but was less common than in the South Atlantic during the same year (Waluda et al., 2002). High levels of fishing have previously been observed in high seas waters off South America, notably when squid were particularly abundant off the coast of Peru, e.g. during 1994 (Waluda & Rodhouse, in press). The commercial fishery for D. gigas is fairly recent, beginning in 1991 (Yamashiro et al., 1998), and is probably not as intensively exploited as the I. argentinus fishery. High seas fishing may be less important in terms of stock conservation, but DMSP-OLS can be used to quantify the distribution and abundance of the fleet fishing off the coast of Peru and provides a means of monitoring fishing outside EEZ waters. This work has allowed us to estimate how much light (in terms of area) is emitted by a single squid fishing vessel, and may prove to be a powerful tool in assessing fisheries using satellite remote sensing. Acknowledgements Support from the British Antarctic Survey (Natural Environmental Research Council) is gratefully acknowledged. We would like to thank two anonymous referees for their helpful comments on an earlier draft. This research forms part of an Independent Project awarded to PGR and funded under the British Antarctic Survey s Antarctic Science in the Global Context Programme. References Anderson, C. I. H., & Rodhouse, P. G. (2001). Life cycles, oceanography and variability: ommastrephid squid in variable oceanographic environments. Fisheries Research, 54, Cho, K., Ito, R., Shimoda, H., & Sakata, T. (1999). Fishing fleet lights and surface temperature distribution observed by DMSP/OLS sensor. International Journal of Remote Sensing, 20, 3 9. Elvidge, C. D., Baugh, K. E., Dietz, J. B., Bland, T., Sutton, P. C., & Kroehl, H. W. (1999). Radiance calibration of DMSP-OLS low light imaging data of human settlements. Remote Sensing of Environment, 68, FAO. (2000). Fishstat plus: Universal software for fishery statistical time series. Rome: FAO Fisheries Department, Fishery Information, Data and Statistics Unit. Ichii, T., Mahapatra, K., Watanabe, T., Yatsu, A., Inagake, D., & Okada, Y. (2002). Occurrence of jumbo flying squid Dosidicus gigas aggregations associated with the countercurrent ridge off the Costa Rica Dome during 1997 El Niño and 1999 La Niña. Marine Ecology Progress Series, 231, Longhurst, A. R. (1998). Ecological geography of the sea ( pp ). San Diego: Academic Press. Nesis, K. N. (1983). Dosidicus gigas. In P. R. Boyle (Ed.), Cephalopod Life Cycles, vol. 1 ( pp ). London: Academic Press. Nigmatullin, C. M., Nesis, K. N., & Arkhipkin, A. I. (2001). A review of the biology of the jumbo squid Dosidicus gigas (Cephalopoda: Ommastrephidae). Fisheries Research, 54, Rodhouse, P. G., Elvidge, C. D., & Trathan, P. N. (2001). Remote sensing of the global light-fishing fleet: an analysis of interactions with oceanography, other fisheries and predators. Advances in Marine Biology, 39, Taipe, A., Yamashiro, C., Mariátegui, L., Rojas, P., & Roque, C. (2001). Distribution and concentrations of jumbo flying squid (Dosidicus gigas) off the Peruvian coast between 1991 and Fisheries Research, 54, Thomas, A. C., Huang, F., Strub, P. T., & James, C. (1994). Comparison of the seasonal and interannual variability of phytoplankton pigment concentrations in the Peru and California current systems. Journal of Geophysical Research, 99, Waluda, C. M., & Rodhouse, P. G., (in press). Dosidicus gigas fishing grounds in the Eastern Pacific as revealed by satellite imagery of the light-fishing fleet. Bulletin of the Phuket Marine Biological Center. Waluda, C. M., Trathan, P. N., Elvidge, C. D., Hobson, V. R., & Rodhouse, P. G. (2002). Throwing light on straddling stocks of Illex argentinus: assessing fishing intensity with satellite imagery. Canadian Journal of Fisheries and Aquatic Science, 59, Yamashiro, C., Mariátegui, L., Rubio, J., Arguelles, J., Tafur, R., Taipe, A., & Rabí, M. (1998). Jumbo flying squid fishery in Peru. In T. Okutani (Ed.), Large pelagic squids ( pp ). Tokyo: Japan Marine Fishery Resources Research Center.