39In the Shiretoko Peninsula, a World Natural Heritage site, the sustainable management of kelp forests has 40 drawn public attention because of the economic and ecological importance of kelp. We spatially estimated
58Kelp forests provide valuable fishery products in Japan. In particular, Laminaria spp. products from the 59 waters of northeastern Hokkaido, Japan, are considered highly valuable because of their superior quality [1].
60Kelp harvesting is an active fisheries industry in Japan, and kelp forests play important ecological roles in 61 the coastal waters [2]. The primary production of kelp forests is as high as that of terrestrial vascular plants,
62such as mature rain forests (approximately 1300 g C m -2 year -1 ), and kelp forests are considered to be major 63 primary producers in coastal ecosystems [3]. Due to these economic and ecological contributions, fisheries
64scientists and environmental managers regard the relationship between kelp distribution and kelp
65harvesting as important information for the sustainability of coastal area.
66The quantification of harvest efforts and identification of harvest locations are key for sustainable 67 fisheries management in coastal area. However, the long coastline has historically made kelp surveys 68 difficult, so information on kelp distribution is limited.
109In this study, we considered that the vessel motion didn't almost affect the amplitude of sea bottom, because
110of the ship speed, the sea condition and the survey depth.
111The sampling equipment used to detect and measure kelp forests consisted of an acoustic component 118 1 (weak) to Lv. 255 (strong). In this study, we set the echo via the PC that operated the onboard system, and
119we saved the intensities in digital data to the PC. Onboard or underwater video observations were 120 conducted to distinguish kelp from other algae (Fig. 2)
148where n c is the number of these pairs, z is the values at locations i and i + h of the detected kelp forest, and 149 h is the lag, which is the distance between pairs of data points at specific locations [20].
150We focused on spatial aspects of kelp forest over the Shiretoko coastal shelf, rather than the dynamics of
9The best-fit theoretical semivariograms were selected from the obtained experimental semivariograms 157 using the maximum likelihood algorithm for before and after harvest (Fig. 4, [20]). A spherical model was
168Kelp forest distribution was analyzed from two different aspects: the presence or absence of forests or 169 forest patches, and the variation in thickness within or among existing forest patches. Here, we describe the 170 horizontal distribution properties of kelp forests and discuss the causal mechanisms of their biological 171 distribution [9, 10, 20]. We applied a two-stage approach to predict the presence or absence of kelp forest 172 and then estimate the thickness of existing forests in each subarea. In the first stage, the probability of kelp 173 occurrence was predicted using probability kriging with the best-fit...