Using cumulative human-impact models to reveal global threat patterns for seahorses

Understanding threats acting on marine organisms and their conservation status is vital but challenging given a paucity of data. We studied the cumulative human impact (CHI) on and conservation status of seahorses (Hippocampus spp.)—a genus of rare and data-poor marine fishes. With expert knowledge and relevant spatial data sets, we built linear-additive models to assess and map the CHI of 12 anthropogenic stressors on 42 seahorse species. We examined the utility of indices of estimated impact (impact of each stressor and CHI) in predicting conservation status for species with random forest (RF) models. The CHI values for threatened species were significantly higher than those for nonthreatened species (category based on International Union for Conservation of Nature Red List). We derived high-accuracy RF models (87% and 96%) that predicted that 5 of the 17 data-deficient species were threatened. Demersal fishing practices with high bycatch and pollution were the best predictors of threat category. Major threat epicenters were in China, Southeast Asia, and Europe. Our results and maps of CHI may help guide global seahorse conservation and indicate that modeling and mapping human impacts can reveal threat patterns and conservation status for data-poor species. We found that for exploring threat patterns of focal species, species-level CHI models are better than existing ecosystem-level CHI models.