Identifying climate refugia for vulnerable marine ecosystem indicator taxa under future climate change scenarios

Vulnerable Marine Ecosystems (VMEs) are recognised as having high ecological significance and susceptibility to disturbances, including climate change. One approach to providing information on the location and biological composition of these ecosystems, especially in difficult-to-reach environments such as the deep sea, is to generate spatial predictions for VME indicator taxa. In this study, the Random Forest algorithm was used to model the spatial distribution of density for 14 deep-water VME indicator taxa under current environmental conditions and future climate change scenarios (SSP2-4.5 and SSP3-7.0) within the New Zealand Territorial Sea and Exclusive Economic Zone (100–1500 m water depth) to evaluate potential changes in the location and distribution of density of these taxa over time. Overall, our species distribution models performed well for all taxa (mean AUC = 0.82; TSS = 0.56; r = 0.40) and predicted a considerable average reduction in density (54%) and habitat extent (61%), by the end of the 21st century under both climate change scenarios. Nevertheless, models identified regions that might serve as internal refugia (approximately 158,000 km 2 ), where some taxa are predicted to maintain the high densities predicted for current-day environmental conditions under future climatic conditions, and external refugia (approximately 121,000 km 2 ) where taxa were predicted to expand into new locations by the end of the 21st century. Our results represent a significant step forward as they provide predictions of the distribution of taxa densities, rather than just occurrence, under both present and future climatic conditions. Furthermore, these findings carry implications for ecosystem management and spatial planning, suggesting current marine spatial protection measures may not offer adequate protection to VME indicator taxa in the face of climate change. Additionally, activities like bottom trawling, present or future, may jeopardize climate refugia viability. Thus, a comprehensive assessment of cumulative effects on VME indicator taxa is recommended to establish effective protection measures for potential climate refugia, ensuring the continuity of essential ecosystem services. Highlights • Density-based SDMs are used to assess the distribution of 14 VME indicator taxa. • Density distribution is projected at both present-day and future climate change scenarios. • Results indicate significant losses in density (54%) and habitat extent (61%) under climate change. • Models identify regions of potential climate refugia. • Predicted refugia areas are currently under low protection and vulnerable to local impacts.