330 NOAA First U.S. Commissioner National Marine Fishery Bulletin established 1881 of Fisheries and founder Fisheries Service of Fishery Bulletin Abstract—Fishing locations for Pa- cific saury (Cololabis saira) obtained Predicting potential fishing zones for Pacific from images of the Operational saury (Cololabis saira) with maximum entropy Linescan System (OLS) of the U.S. Defense Meteorological Satellite models and remotely sensed data Program, together with maximum entropy models and satellite-based 1,2 oceanographic data of chlorophyll- Achmad F. Syah (contact author) a concentration (chl-a), sea-surface Sei-Ichi Saitoh1,3 temperature (SST), eddy kinetic en- Irene D. Alabia3 ergy (EKE), and sea-surface height 1 anomaly (SSHA), were used to evalu- Toru Hirawake ate the effects of oceanographic con- ditions on the formation of potential Email address for contact author: [email protected] fishing zones (PFZ) for Pacific saury and to explore the spatial variabil- 1 Laboratory of Marine Environment and Resource Sensing ity of these features in the western Faculty of Fisheries Sciences North Pacific. Actual fishing regions Hokkaido University were identified as the bright areas 3-1-1 Minato-cho created by a 2-level slicing method Hakodate 041-8611, Japan for OLS images collected August–De- 2 Department of Marine Science cember during 2005–2013. The re- University of Trunojoyo Madura sults from a Maxent model revealed Jalan Raya Telang its potential for predicting the spa- P.O. Box 2 Kamal tial distribution of Pacific saury and Bangkalan-Madura, Indonesia highlight the use of multispectral 3 Arctic Research Center satellite images for describing PFZs. Hokkaido University In all monthly models, the spatial N21-W11 Kita-ku PFZ patterns were explained pre- Sapporo 001-002, Japan dominantly by SST (14–16°C) and indicated that SST is the most influ- ential factor in the geographic distri- bution of Pacific saury. Also related to PFZ formation were EKE and SSHA, possibly through their effects on the feeding grounds conditions. The Pacific saury (Cololabis saira) environmental factors on abundance Concentration of chl-a had the least is widely distributed in the west- of Pacific saury was evident in the effect among other environmental ern North Pacific from subarctic to unexpected drop in both the catch factors in defining PFZs, especially during the end of the fishing season. subtropical waters and is one of the and catch per unit of effort in 1998, commercially important pelagic spe- following a period of high abundance cies in Japan, Russia, Korea, and (Tian et al., 2003). The distribution Taiwan. The total landings of this and migratory patterns of Pacific species in these countries increased saury have been associated with from 171,692 metric tons (t) in 1998 chlorophyll-a (chl-a) concentration to 449,738 t in 2011. Over the last and sea-surface temperature (SST) half century, annual catches of Pacif- (Watanabe et al., 2006; Mukai et ic saury in Japan, for example, have al., 2007; Tseng et al., 2013). More- averaged around 257,800 t (Tian et over, sea-surface height indicates Manuscript submitted 20 July 2015. al., 2003) and have fluctuated greatly water mass movements and, by ex- Manuscript accepted 12 May 2016. from 52,207 t in 1969 to 207,770 t in tension, the flow of heat and nutri- Fish. Bull.:330–342 (2016). 2011 (Fisheries Agency and Fisheries ents, which will subsequently influ- Online publication date: 2 June 2016. Research Agency of Japan, 2012). ence productivity (Ayers and Lozier, doi: 10.7755/FB.114.3.6 The number, size, and location of 2010). Sea-surface height can also be The views and opinions expressed or fishing grounds for Pacific saury are used to infer physical oceanographic implied in this article are those of the largely affected by oceanographic features, such as eddies, fronts, and author (or authors) and do not necessarily conditions (Yasuda and Watanabe, convergences (Polovina and Howell, reflect the position of the National 1994; Kosaka, 2000; Tian et al., 2005). Therefore, understanding the Marine Fisheries Service, NOAA. 2002), and the significant effect of relationship between oceanographic Syah et al.: Predicting potential fishing zones for Cololabis saira 331 factors and the migration and distribution of species is Materials and methods essential for fisheries management. Most studies of Pacific saury have concentrated on Study area distribution and migration and have used in situ or logbook data (Huang et al., 2007; Tseng et al., 2013), This study was conducted in the western North Pacific, and models have been developed to investigate growth extending from 140° to 155°E and from 34° to 46°N and abundance (Tian et al., 2004; Ito et al., 2004, 2007; (Fig. 1). In this study area, located between the sub- Mukaietal., 2007). In contrast, Watanabe et al. (2006) arctic and subtropical domains of the North Pacific, proposed a spatial and temporal migration model for the confluence of the warm Kuroshio Current and the stock size that was dependent on SST. However, inte- cold Oyashio Current forms the Kuroshio–Oyashio grated high-resolution nighttime satellite images, such transition zone (Roden, 1991), also called the subarc- as those available in the time-series data from the Op- tic–subtropical transition zone. The Kuroshio Current erational Linescan System (OLS) of the Defense Me- is characterized by warm, low-density, nutrient-poor, teorological Satellite Program, U.S. Department of De- and high-salinity surface waters (Yatsu et al., 2013), fense, together with habitat and environmental model- whereas the Oyashio Current is characterized by low- ing, have not been used to predict the potential fishing salinity, low-temperature, and nutrient-rich waters zones for Pacific saury. (Sakurai, 2007). The Kuroshio–Oyashio transition In Japan, fishing vessels operate at night and use zone is characterized by the mixing of various water stick-held dip nets, locally known as bouke ami, which masses and complex physical oceanographic structures are equipped with lights to attract fishes (Fukushima, (Roden, 1991). Moreover, 3 major oceanic fronts exist 1979). These fishing vessels, equipped with lights, as in this region: the Polar Front, Subarctic Front, and are vessels that fish for Pacific saury, can be identified Kuroshio Extension Front (Science Council of Japan1). by the OLS sensor, which also enables the detection The characteristic patterns of these oceanic fronts also of moonlight-illuminated clouds and lights from cit- have been well documented in earlier studies (Kitano, ies, towns, industrial sites, gas flares, and ephemeral 1972; Roden et al., 1982; Belkin and Mikhailichenko, events, such as fires and lightning-illuminated clouds 1986; Miyake, 1989; Belkin et al., 1992, 2002; Yoshida, (Elvidge et al., 1997). In addition, OLS nighttime im- 1993; Onishi, 2001; Murase et al., 2014; Shotwell et ages can be used to estimate fishing vessel numbers al., 2014). and fishing areas for squid (Kiyofuji and Saitoh, 2004; Kiyofuji et al., 2004). The relationship between the Satellite nighttime images number of lit pixels in OLS nighttime images and the number of fishing vessels also has been analyzed for Daily cloud-free OLS nighttime images were download- the fishery of Illex argentinus (Waluda et al., 2002). ed from the Satellite Image Database System of the The brightly lit areas seen in nighttime images of the Agriculture, Forestry and Fisheries Research Informa- western North Pacific are the result of vessels fishing tion Center of the Japan Ministry of Agriculture, For- for Pacific saury or squid (Semedi et al., 2002; Saitoh estry and Fisheries [the system is no longer operating]. et al., 2010; Mugo et al., 2014). The images were then used to determine the location Predictive habitat modeling has become an increas- of the vessels that fish for Pacific saury in the western ingly useful tool for marine ecologists and conservation North Pacific. A TeraScan2 system, vers. 4.0 (Seaspace scientists in order to estimate the patterns of species Corp., Poway, CA) was used to analyze the images and distribution and to subsequently develop conservation to process the nighttime lights into digital numbers strategies (Johnson and Gillingham, 2005; Tsoar et (DNs), in a range of 0–63, that represent the visible al., 2007; Ready et al., 2010). The maximum entropy pixels in relative values. We selected 1264 single pass method (Phillips et al., 2006) involves one of the most images collected from August through December dur- widely used machine-learning algorithms for inferring ing 2005–2013 (9 years) by 6 Defense Meteorological species distributions. In recent studies, the method of Satellite Program satellites (F13, F14, F15, F16, F17, maximum entropy has been applied to both terrestrial and F18) (Table 1). The period from August through (Peterson et al., 2007) and marine ecosystems (Ready December was chosen for analysis because it corre- et al., 2010; Edrén et al., 2010; Alabia et al., 2015). In sponds with the fishing season of Pacific saury. To con- this study, we used a maximum entropy approach with struct the habitat suitability model, the daily images multi sensor satellite datasets and OLS-derived spe- were reprocessed with a 1-km resolution and then com- cies occurrences to create an accurate prediction model piled in a monthly database. The location of the vessels and investigate the potential fishing zones for Pacific was assumed to represent the location of Pacific saury. saury in the western North Pacific. The objectives of this study were to evaluate the effects of oceanographic 1 Science Council of Japan. 1960. The results of the Japa- factors on the formation of potential fishing zones for nese oceanographic project for the International Geophysical Pacific saury and to examine the variability in spatial Year 1957/8, 145 p. National Committee for the Interna- tional Geophysical Year, Science Council of Japan, Tokyo. patterns of potential fishing zones in relation to the 2 Mention of trades names or commercial companies is for prevailing oceanographic conditions in the western identification purposes only and does not imply endorsement North Pacific.