Benthic habitats are primarily determined by the nature of the substrate (i.e., the type of sediment or rock) and by the bathymetry (e.g., light availability, current, or waves exposition) both determining the viability of infaunal organisms. In this sense multibeam (MB) echosounders provide more data about the sea bottom than any other acoustic technique. During the 2014 Argentine-Canadian survey of the RV Coriolis II in the Gulf of San Jorge, MB data were acquired in the area of Robredo, west of the Marine Protected Area (MPA) Parque Interjurisdiccional Marino Costero Patagonia Austral (PIMCPA, Patagonia, Argentine). Acoustic data were affected by acquisition artifacts, so an ad hoc algorithm was required to restore the data. Moreover, during the survey, dredge samples were also acquired where acoustic data suggested changes in substrate properties, and two later scientific surveys acquired video transects of the sea bottom that allowed also a basic classification. A non-supervised classification of the seabed was performed, using a subsampling and voting algorithm, with the acoustic and bathymetric data. The final classification is in agreement with groundtruthing (dredge samples and video transects), showing that bottom structure complexity leads to a meaningful picture of the sea bottom classes in the area. Our results reveal that Robredo has two subareas, one with a fine substrate (west) and another with a coarse substrate (east), the latter being also characterized by outcropping rock formations of marked relief. Despite the small spatial coverage of the survey, the resulting sea bottom classification presented in this paper has described, for the first time, some important features of the benthic geohabitats in the PIMCPA.

底栖生境主要由基质的性质（即沉积物或岩石的类型）和测深法（例如光的可利用性，水流或海浪的暴露）确定，它们均决定着不孕生物的生存力。从这个意义上讲，多波束（MB）回声测深仪比任何其他声学技术都能提供更多有关海底的数据。在2014年阿根廷和加拿大对圣豪尔赫湾RV科里奥利二世的调查中，在海洋保护区（MPA）以西国家海洋保护区Marino Costero Patagonia Austral以西的罗布雷多地区获得了MB数据（PIMCPA，巴塔哥尼亚，阿根廷）。声音数据会受到采集伪影的影响，因此需要一种临时算法来恢复数据。此外，在调查期间，还采集了挖泥机样本，这些样本的声学数据表明底材特性发生了变化，随后的两次科学调查也采集了海底的视频剖面，这也使得基本分类成为可能。使用子采样和投票算法，对声学和测深数据进行了海底非监督分类。最终的分类与地面挖泥（挖泥船样本和视频样带）相吻合，表明底部结构的复杂性导致了该地区海底类别的有意义的描述。我们的结果表明，罗布雷多有两个分区，一个具有良好的基质（西部），另一个具有粗糙的基质（东部），后者的特征还在于露头明显的岩层。尽管调查的空间覆盖范围很小，但本文中提出的最终海底分类还是首次描述了PIMCPA中底栖地理栖息地的一些重要特征。