AUTHOR=Matsubayashi Jun , Kimura Katsuya , Ohkouchi Naohiko , Ogawa Nanako O. , Ishikawa Naoto F. , Chikaraishi Yoshito , Tsuda Yuichi , Minami Hiroshi 

TITLE=Using geostatistical analysis for simultaneous estimation of isoscapes and ontogenetic shifts in isotope ratios of highly migratory marine fish

JOURNAL=Frontiers in Marine Science

VOLUME=9

YEAR=2022

URL=https://www.frontiersin.org/journals/marine-science/articles/10.3389/fmars.2022.1049056

DOI=10.3389/fmars.2022.1049056

ISSN=2296-7745

ABSTRACT=<p>Tracking migration of highly migratory marine fish using isotope analysis (iso-logging) has become a promising tool in recent years. However, application of this method is often hampered by the lack of essential information such as spatial variations in isotope ratios across habitats (isoscapes) and ontogenetic shifts of isotope ratios of target animals. Here, we test the utility of geostatistical analysis to generate isoscapes of <italic>δ</italic><sup>13</sup>C and <italic>δ</italic><sup>15</sup>N in the western Pacific and estimate the ontogenetic shifts in <italic>δ</italic><sup>13</sup>C and <italic>δ</italic><sup>15</sup>N values of a target species. We first measured <italic>δ</italic><sup>13</sup>C and <italic>δ</italic><sup>15</sup>N in the white muscle of juvenile (<italic>n</italic> = 210) and adult (<italic>n</italic> = 884) skipjack tuna <italic>Katsuwonus pelamis</italic> sampled across the northwest Pacific. Next we fitted a geostatistical model to account for the observed spatial variations in <italic>δ</italic><sup>13</sup>C and <italic>δ</italic><sup>15</sup>N of skipjack by fork length and other environmental variables with spatial random effects. We then used the best-fit models to predict the isoscapes of <italic>δ</italic><sup>13</sup>C and <italic>δ</italic><sup>15</sup>N in 2021. Furthermore, we measured <italic>δ</italic><sup>15</sup>N of amino acids (<italic>δ</italic><sup>15</sup>N<sub>AAs</sub>) of skipjack (<italic>n</italic> = 5) to determine whether the observed spatial variation of isotope ratios resulted from baseline shifts or differences in trophic position. The geostatistical model reasonably estimated both isoscapes and ontogenetic shifts from isotope ratios of skipjack, and the isoscapes showed that <italic>δ</italic><sup>13</sup>C and <italic>δ</italic><sup>15</sup>N can clearly distinguish the latitudinal migration of skipjack in the western Pacific. The <italic>δ</italic><sup>15</sup>N<sub>AAs</sub> supported the results of the geostatistical model, that is, observed variations in skipjack <italic>δ</italic><sup>15</sup>N were largely derived from a baseline shift rather than regional differences in trophic position. Thus, we showed that geostatistical analysis can provide essential basic information required for iso-logging without compound-specific isotope analysis.</p>