Main content area

Acknowledging differences: number, characteristics, and distribution of marine benthic communities along Taiwan coast

Lin, Yuting Vicky, Denis, Vianney
Ecosphere 2019 v.10 no.7 pp. e02803
algorithms, artificial intelligence, benthic organisms, coasts, coral reefs, endangered ecosystems, photoperiod, surface water temperature, variance, Taiwan
Our knowledge of the ability of coral reefs to face contemporary challenges disregards the wide variety of benthic communities (BCs) that could be associated with contrasted dynamics when facing stressors. Accordingly, this study investigates the number, characteristics, and distribution of BCs responsible for the development of coastal three‐dimensional structures in Taiwan. A total of 89 transects among three regions (north, east, and south) and two depths (−10 and −40 m) were characterized using a morpho‐functional categorization of benthic organisms. Using an unsupervised learning algorithm, k‐means cluster, an optimal number of k groups were identified among transects in order to minimize total intra‐group variance and represent a first level of organization, mirroring the number of BCs. Each BC was then described into prevailing categories and typified by an association with significant indicator groups. Their distributions were further examined and tested among regions and depths. Seven BCs were identified as having different composition and indicator categories. Their distributions suggest that, while sea surface temperature and light regime may be associated with a panel of available BCs, local conditions may have the final say on whether a BC eventually thrives at a given location. Overall, this study proposes an innovative and simple analytical framework for acknowledging differences among BCs. Our results encourage greater consideration of these shades in coral reef ecology studies, as they could conceal a variety of roles and dynamics important for the conservation of these endangered ecosystems.