The ecological integrity of riverine systems is influenced both by environmental conditions and anthropogenic modifications. The Moselle river, a tributary of the Rhine has undergone significant structural changes for decades. This paper investigates the effects of these changes on the localizations of the species that are found in the Barbel zone of the river. We employ a deep auto-encoder based clustering approach to first cluster the species found. Then each cluster is separately analyzed and then links the findings to the changing environmental conditions alongside modifications in anthropogenic conditions. Our results highlight distinct clustering patterns that reflect the combined effects of habitat conditions, flow velocity of water columns in certain stations, presence of navigation channels and substrate characteristics on fish communities. These findings provide us with insights regarding the spatial organization of the diversity of fish for future ecosystem management.

错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

Deep Auto-encoder Learning of Fish Species Clustering in the Upstream Sectors of the Moselle River

  • Samudranil Basak,
  • Gerard Masson,
  • Hoai Minh Le,
  • Baba Issa Camara

摘要

The ecological integrity of riverine systems is influenced both by environmental conditions and anthropogenic modifications. The Moselle river, a tributary of the Rhine has undergone significant structural changes for decades. This paper investigates the effects of these changes on the localizations of the species that are found in the Barbel zone of the river. We employ a deep auto-encoder based clustering approach to first cluster the species found. Then each cluster is separately analyzed and then links the findings to the changing environmental conditions alongside modifications in anthropogenic conditions. Our results highlight distinct clustering patterns that reflect the combined effects of habitat conditions, flow velocity of water columns in certain stations, presence of navigation channels and substrate characteristics on fish communities. These findings provide us with insights regarding the spatial organization of the diversity of fish for future ecosystem management.