ARCTIC MARINE FOREST DISTRIBUTION MODELS SHOWCASE SEVERE NET LOSSES AND TEMPERATE SUCCESSION UNDER CLIMATE CHANGE

• Main Author: Bringloe, Trevor
• Format: Dataset
• Language: unknown
• Published: figshare 2021
• Subjects: 69902 Global Change Biology; FOS Biological sciences; Arctic; Climate change
• Online Access: https://dx.doi.org/10.6084/m9.figshare.14751753.v4; https://figshare.com/articles/dataset/ARCTIC_MARINE_FOREST_DISTRIBUTION_MODELS_SHOWCASE_SEVERE_NET_LOSSES_AND_TEMPERATE_SUCCESSION_UNDER_CLIMATE_CHANGE/14751753/4

Our main objective was to investigate potential gains and losses in northern hemisphere marine forest habitat along temperate and Arctic coastlines under different climate change scenarios. In particular, we sought to investigate anticipated responses according to three ecosystem distribution categories: marine forests restricted to the Arctic (i.e. oceanic benthic waters with a mean average annual temperature of ca. <5oC), widely distributed species occurring in temperate and Arctic waters, and species restricted to cold and warm temperate waters with no current range in the Arctic. The datasets provided here include the occurrence records for training models, area calculations for both threshold and probabilistic predictions for each species and each climate change scenario, and the maxent predictions for each species, including the stacked predictions, and for each climate change scenario. v2 provides updated models using long term min and max benthic temperature layers, which was done to better reflect longer term temperature trends. The original models (which used absolute min and max benthic temperature layers) are provided in the zip folder "old_models_absolute_min_max_temp.zip". Also included are the accompanying publication figures for the old models. Updated figures will be featured in the final publication of this work. Model performance metrics and MaxEnt output (including response curves) are also provided in v2. For threshold predictions, yellow=occurrence, blue=no occurrence. For probabilistic predictions, pink->blue->green->orange->red refers to 1.0->0 probability of occurrence.