Subpolar North Atlantic Ocean model biases and trends in ocean-sea ice models driven by historic atmospheric forcing

The state of the North Atlantic Ocean is highly scrutinized because of the potential for an Atlantic meridional overturning circulation (AMOC) slowdown under anthropogenic forcing. Such a trend, which is projected by global climate models (GCMs), has implications for climate in the North Atlantic and downstream. GCMs, however, have longstanding biases in North Atlantic Ocean temperature, salinity, and currents, many of which can be attributed to the ocean model. Such issues call into question the validity of global ocean model representation of the AMOC. Are such biases severe enough to impact the representation of historic trends in North Atlantic thermohaline circulation? To address this question, we analyze North Atlantic Ocean model biases and trends using interannual forced simulations from the Ocean Model Intercomparison Project (OMIP). Interannual forced OMIP simulations are integrated from the mid-twentieth century to near-present and are driven by atmospheric forcing derived from reanalyses. We quantify OMIP biases of surface and upper-ocean temperature and salinity, and consider the multi-model range of the overturning circulation. To understand how biased surface forcing impacts modeled overturning circulation, we examine surface-forced water mass transformation and compare it to overturning streamfunction in density space. A new observation-based benchmark is used to identify transformation biases. By comparing to observations or observation-based benchmarks where possible, we will consider the effect that North Atlantic Ocean thermohaline biases may have on historic AMOC trends.