#More phone numbers: https://tel.meet/wde-shmw-vcw?pin=3526342705011

Abstract: Deep-sea coral distributions in the Northwest Atlantic remain poorly characterized due to sampling limitations. We developed species distribution models (SDMs) for the cold-water coral Desmophyllum dianthus using generalized additive models (GAMs) to identify key environmental drivers and predict suitable habitat. We analyzed 81,112 presence-absence records from multiple research cruises, evaluating 28 environmental predictors through variance inflation factor (VIF) screening and univariate assessment. The final model incorporated depth, rugosity index (rie), slope (qslp), and bottom total alkalinity (btm_talk_ann) as smoothed terms, with and without spatial coordinates. Model performance was strong (AUC = 0.878 with location, 0.849without), with 26.0% deviance explained. D. dianthus showed a unimodal depth response peaking at 500 ” 700 m, positive associations with seafloor rugosity and slope, and a narrow alkalinity optimum (2.400 ” 2.405 mol m-3). Spatial structure accounted for 18.5% of explained deviance, suggesting unmeasured environmental gradients or dispersal limitations contribute moderately to distribution patterns. The 4.8 percentage point deviance improvement and 0.029AUC increase when including spatial terms demonstrates the value of incorporating geographic structure in deep-sea SDMs, though environmental predictors remain primary drivers. Our results provide critical baseline information for conservation planning and highlight the importance of topographic complexity and oceanographic conditions in determining cold-water coral distributions. The results are from the NOAA EPP/MSI CSC NERTO graduate internship project, which was conducted under the guidance of NOAA mentor James Vasslides of James J. Howard Marine Sciences Laboratory at Sandy Hook. The NERTO aligns with NOAA CSC Center for Earth System Sciences and Remote Sensing Technologies (CESSRST-II) award’s goal of becoming a scientist. The NERTO also deepened the intern’s understanding of multiple modeling approaches that include generalized linear models (GLMs), generalized additive models(GAMs), boosted regression trees (BRTs), and random forest models (RFs), using has been collected as part of the larger Northeast Deep Sea Coral Initiative, he will interact with scientists from other parts of NOAA Fisheries, NCCOS, and international collaborators.