Anthropogenic climate change is predicted to decrease oceanic oxygen (O2) concentrations, with potentially significant effects on marine ecosystems. Geologically recent episodes of abrupt climatic warming provide opportunities to assess the effects of changing oxygenation on marine communities. Thus far, this knowledge has been largely restricted to investigations using Foraminifera, with little being known about ecosystem-scale responses to abrupt, climate-forced deoxygenation.

European continental shelf seas have experienced intense warming over the past 30 years. In the North Sea, fish have been comprehensively monitored throughout this period and resulting data provide a unique record of changes in distribution and abundance in response to climate change. We use these data to demonstrate the remarkable power of generalized additive models (GAMs), trained on data earlier in the time series, to reliably predict trends in distribution and abundance in later years.

This paper presents key results from analysis of surface meteorological observations collected in the Northern Arabian/Persian Gulf (N Gulf; Kuwait, Bahrain, and NE Saudi Arabia), which spans a 40-years period (1973–2012). The first part of this study analyzes climate variability in the N Gulf, and relates them to teleconnection patterns (North Atlantic Oscillation, El Nino Southern Oscillation, and Indian Ocean Dipole).

In a recent letter, Ricke and Caldeira (2014 Environ. Res. Lett. 9 124002) estimated that the timing between an emission and the maximum temperature response is a decade on average. In their analysis, they took into account uncertainties about the carbon cycle, the rate of ocean heat uptake and the climate sensitivity but did not consider one important uncertainty: the size of the emission. Using simulations with an Earth System Model we show that the time lag between a carbon dioxide (CO2) emission pulse and the maximum warming increases for larger pulses.