Predicting the effects of climate change on deep‐water coral distribution around New Zealand—Will there be suitable refuges for protection at the end of the 21st century?

Abstract

Deep-water corals are protected in the seas around New Zealand by legislation that prohibits intentional damage and removal, and by marine protected areas where bottom trawling is prohibited. However, these measures do not protect them from the impacts of a changing climate and ocean acidification. To enable adequate future protection from these threats we require knowledge of the present distribution of corals and the environmental conditions that determine their preferred habitat, as well as the likely future changes in these conditions, so that we can identify areas for potential refugia. In this study, we built habitat suitability models for 12 taxa of deep-water corals using a comprehensive set of sample data and predicted present and future seafloor environmental conditions from an earth system model specifically tailored for the South Pacific. These models predicted that for most taxa there will be substantial shifts in the location of the most suitable habitat and decreases in the area of such habitat by the end of the 21st century, driven primarily by decreases in seafloor oxygen concentrations, shoaling of aragonite and calcite saturation horizons, and increases in nitrogen concentrations. The current network of protected areas in the region appear to provide little protection for most coral taxa, as there is little overlap with areas of highest habitat suitability, either in the present or the future. We recommend an urgent re-examination of the spatial distribution of protected areas for deep-water corals in the region, utilising spatial planning software that can balance protection requirements against value from fishing and mineral resources, take into account the current status of the coral habitats after decades of bottom trawling, and consider connectivity pathways for colonisation of corals into potential refugia.

Modelling spatial distributions of biogenic habitat‐forming taxa to inform marine spatial planning

“…the combination of outputs from more than one model type) were generated from boosted regression tree (BRT) and random forest (RF) models. These approaches were chosen following previous studies that showed they perform well for similar modelling tasks in the same region (Anderson et al, 2022; Finucci et al, 2021; Stephenson et al, 2023). The modelling approach used here closely follows the method described in Stephenson et al (2023).…”

Using joint species distribution modelling to predict distributions of seafloor taxa and identify vulnerable marine ecosystems in New Zealand waters

“…Following development of JSDMs for benthic taxa for the entire New Zealand EEZ, and on the completion of the further sampling and analyses suggested above, it may be possible to quantitatively incorporate robust predictions of VMEs into future spatial planning efforts in the region. Ideally, this planning would also include other anthropogenic threats to the biodiversity supported by VMEs, including climate change effects (e.g., as in Anderson et al, 2022). Achieving these steps would provide a wealth of additional information on sea oor communities and habitats for use in future spatial planning efforts.…”

Key factors for species distribution modeling in benthic marine environments

“…The selection of background data for modeling in this study used commonly employed prevalence ratios of 1:1, 1:5, and 1:10, as well as a fixed number of 10,000 points (Phillips et al, 2006;Barbet-Massin et al, 2012;Hysen et al, 2022). Two methods were used to select background points: the target-group background sampling method and kernel density sampling method (Elith et al, 2010;Fitzpatrick et al, 2013;Cerasoli et al, 2017;Georgian et al, 2019;Burgos et al, 2020;Finucci et al, 2021;Georgian et al, 2021;Robinson et al, 2021;Stephenson et al, 2021;Anderson et al, 2022). To select target-group backgrounds for each species, random points were selected from the remaining cold-water coral presences with the selected points within 5 km of occurrences excluded.…”

“…The target-group sampling method utilizes all occurrences within a target group as biased background data, thereby selecting the background points with the same bias as the sampling effort to reduce the impact of sampling bias in presence-background modeling (Phillips et al, 2009;Merow et al, 2013). Target-group sampling has been found to improve the average performances of tested modeling techniques compared to using randomly selected background data (Phillips et al, 2009;Iturbide et al, 2015), which has been widely used in species distribution predictions in recent years (Cerasoli et al, 2017;Stephenson et al, 2020;Robinson et al, 2021;Stephenson et al, 2021;Anderson et al, 2022).…”