Stability of marine communities in response to climate change

Abstract

This work presents an intuitive method for studying the stability of fish communities in response to climate change via a three-step approach: (i) quantification of climate change across different latitudes, (ii) quantification of community instability, and (iii) quantification of the relationship between climate change and community instability at each location. Abundance data and other biological and environmental variables were recorded for groundfish trawl surveys conducted in the East Bearing Sea during from 1982 to 2011. First, we define climate change as the relationship between (surface, bottom) temperature and time at each location. A significant linear relationship between latitude and bottom water temperature suggests that a change of 1 degree in latitude towards the North Pole leads to an increase in bottom water temperature of 0.035 degrees Celsius. Second, we propose to measure community instability through the computation of three community instability metrics: pairwise distance between observation points/site scores, mean distance to the centroid point, and the area of a fitted convex hull containing all observation points. We compute instability metrics from a low-dimensional representation of the original data, which is built using a non-metric multidimensional scaling technique. Once the instability metrics are calculated, we regress them as a function of latitude to study how the stability of communities changes across locations. We found that all three instability metrics increased with latitude. This trend could be due to the fact that increases in bottom temperature destabilize ecological communities by causing species to migrate to locations with more stable environmental conditions. Finally, with the intuition of how latitude drives the behavior of community stability, we regress the instability metrics as a function of the temporal change in temperature in order to find the environmental factors that are being "proxied" by latitude in the former relationship. We show that towards northern latitudes, where community instability is higher, temporal changes in temperature variables suggest an inverse relationship, with faster temporal changes in the surface water temperatures and slower temporal changes in the bottom water temperatures.