Thresholds of seascape fauna composition along gradients of human pressures and natural conditions to inform marine spatial planning.

Knowledge about the cumulative impacts of anthropogenic activities and environmental conditions on marine ecosystems is incomplete and details are lacking. Compositional community changes can occur along gradients, and community data can be used to assess the state of community resilience against combined impacts of variables representing human pressures and environmental conditions. Here we use a machine learning approach, i.e., Gradient Forest, to identify explanatory variable thresholds and select relevant epibenthic fauna and demersal fish species, which can be used to inform an integrated management of multiple human pressures and conservation planning in the southern North Sea. We show that a broad selection of anthropogenic and environmental variables, such as natural disturbance of the seafloor and euphotic depth, determined community composition thresholds of 67 epibenthic fauna and 39 demersal fish species along environmental conditions and human pressure gradients in the southern North Sea between 2010 and 2020. This has the potential to inform resilience assessments under the Marine Strategy Framework Directive to promote and retain a good environmental status of marine ecosystems.

A risk-based approach to cumulative effects assessment for large marine ecosystems to support transboundary marine spatial planning: A case study of the yellow sea.

Cumulative effects assessment (CEA) should be conducted at ecologically meaningful scales such as large marine ecosystems to halt further ocean degradation caused by anthropogenic pressures and facilitate ecosystem-based management such as transboundary marine spatial planning (MSP). However, few studies exist at large marine ecosystems scale, especially in the West Pacific seas, where countries have different MSP processes yet transboundary cooperation is paramount. Thus, a step-wise CEA would be informative to help bordering countries set a common goal. Building on the risk-based CEA framework, we decomposed CEA into risk identification and spatially-explicit risk analysis and applied it to the Yellow Sea Large Marine Ecosystem (YSLME), aiming to understand the most influential cause-effect pathways and risk distribution pattern. The results showed that (1) seven human activities including port, mariculture, fishing, industry and urban development, shipping, energy, and coastal defence, and three pressures including physical loss of seabed, input of hazardous substances, nitrogen, and phosphorus enrichment were the leading causes of environmental problems in the YSLME; (2) benthic organisms, fishes, algae, tidal flats, seabirds, and marine mammals were the most vulnerable ecosystem components on which cumulative effects acted; (3) areas with relatively high risk mainly concentrated on nearshore zones, especially Shandong, Liaoning, and northern Jiangsu, while coastal bays of South Korea also witnessed high risk; (4) certain risks could be observed in the transboundary area, of which the causes were the pervasive fishing, shipping, and sinking of pollutants in this area due to the cyclonic circulation and fine-grained sediments. In future transboundary cooperation on MSP, risk criteria and evaluation of existing management measures should be incorporated to determine whether the identified risk has exceeded the acceptable level and identify the next step of cooperation. Our study presents an example of CEA at large marine ecosystems scale and provides a reference to other large marine ecosystems in the West Pacific and elsewhere.