Abstract
The optimal spatial resolution for analysis of the spatial distribution of organisms is determined by the scale of their ecological processes (e.g., mobility of target species). This study examined the spatial extent of habitats of five megafaunal species with different levels of mobility [(1) the shrimp, Nematocarcinus sp. (benthic, comparatively low mobility); (2) the galatheid crab, Shinkaia crosnieri (benthic, low mobility); (3) the lithodid crab, Paralomis sp. (benthic, high mobility); (4) the snailfish, Liparidae sp. (demersal, high mobility); and (5) the deep-sea spiny eel, Notacanthidae sp. (demersal, high mobility)], which are associated with a deep-sea hydrothermal field in the Okinawa Trough. We obtained species presence–absence and abiotic environmental data (water temperature, proportional coverage of soft-sediment area, water depth, and five bathymetric parameters) from observation records by deep-sea submersibles and conducted a spatial analysis using three types of statistical models and a multi-model ensemble forecasting method. We prepared 10 data sets of the environmental variables by changing the spatial resolution from finer 0.00005° (ca. 5.3 m) to coarser 0.0005° (ca. 52.5 m). Accuracy of the models for the benthic species with comparatively low mobility was higher than that for the benthic and demersal species with high mobility and enhanced with decreasing grid-cell size (finer resolutions) of environmental data. By contrast, the accuracy of the models for the latter group varied more gently with the changing spatial resolution. The results showed that the proposed method is potentially useful to map the local-scale spatial distribution of the target species, sedentary benthic animals in particular, and that considering an appropriate spatial resolution in predicting the species distributions is important. These findings will help to enable management and subsequent conservation of deep-sea ecosystems.
