Increasing anthropogenic activities are threatening freshwater ecosystems worldwide. Sand mining in Poyang Lake has significantly impacted the wetland ecosystem over the past 20 years, yet a quantitative analysis of these impacts remains insufficient. Here, we used 63 Landsat images taken from 2000 to 2020 along with the support vector machine (SVM) method and a retrieval model of suspended sediment concentration (SSC) to identify sand mining vessels and areas affected by sand mining. Then, we analyzed the changes in landscape patterns in the areas affected by sand mining. The potential impact of underwater noise generated by sand mining vessels on Yangtze finless porpoises was analyzed by a sound propagation model. The number of sand mining vessels in Poyang Lake during the flood, normal, and dry seasons increased from 2000 to 2016 but rapidly decreased from 2017 to 2020. Sand mining vessels were mainly distributed in the northern channel from 2000 to 2006, moved toward the center of the lake from 2007 to 2010, then moved northward in 2017. Within the areas affected by sand mining, water and mudflats declined, grassland and sandbars increased, and the landscape discontinuity increased. The habitat of the Yangtze finless porpoise affected by underwater noise from sand mining vessels in all seasons has significantly increased overtime. The mean area of the affected habitats was 70.65% (dry), 64.48% (normal), and 63.30% (flood) of the total habitat areas. The porpoise habitats in the northern channel and the west branch of the Ganjiang River are more seriously affected by the underwater noise of sand mining vessels than the southern lake. The impact of sand mining activities on wetland landscape and aquatic species demands systematic investigation in the future.

Freshwater cetaceans that inhabit river basins with seasonally changing hydrological regimes have a higher risk of stranding, leading to increased mortality and population decline. In Poyang Lake, the stranding risk of the critically endangered Yangtze finless porpoise are high, due to the significant differences in hydrological and landscape conditions between the flood and dry seasons. However, this stranding information is not well recorded, resulting in poorly guided investigation and rescue efforts. We here employ remote sensing and hydrodynamic modeling to obtain four evaluation indicators in dry, normal, and flood scenarios in Poyang Lake. Results show that nearly 50% of the largest habitat range of the porpoises in the flood season will be land areas during the dry season, and that landscape fragmentation between land and water has increased over the past three decades. In all scenarios, the mean water depth of the habitat varied from 1.77 to 4.89 m from September–February. In the dry scenario, about 59% of the habitat experiences a water depth of 251.04 km² in the normal scenario, and >90.12 km² in the flood scenario. High-stranding risk areas are located within city boundaries, where porpoise stranding cases have been recorded, especially in Duchang, which has the most high-risk areas in all scenarios. In all scenarios, the high stranding risk area has an average bathymetry of 7.81 m and an average water depth between 1.75 and 5.54 m. Our results can guide future investigations to establish sound stranding networks, and the methods proposed here are also applicable to studies of other freshwater cetaceans facing severe stranding risk.