Oceanic dominance of interannual subtropical North Atlantic heat content variability

Abstract

Abstract. Ocean heat content varies on a range of timescales. Traditionally the atmosphere is seen to dominate the oceanic heat content variability. However, this variability can be driven either by oceanic or atmospheric heat fluxes. To diagnose the relative contributions and respective timescales, this study uses a box model forced with output from an ocean general circulation model (OGCM) to investigate the heat content variability of the upper 800 m of the subtropical North Atlantic from 26° N to 36° N. The ocean and air-sea heat flux data needed to force the box model is taken from a 19 yr (1988 to 2006) simulation performed with the 1/12° version of the OCCAM OGCM. The box model heat content is compared to the corresponding heat content in OCCAM for verification. The main goal of the study is to identify to what extent the seasonal to interannual ocean heat content variability is of atmospheric or oceanic origin. To this end, the box model is subjected to a range of scenarios forced either with the full (detrended) ocean and air-sea fluxes, or their deseasoned counterparts. Results show that in all cases, the seasonal variability is dominated by the seasonal component of the air-sea fluxes, which produce a seasonal range in mean temperature of the upper 800 m of ~ 0.42 °C. However, on longer timescales oceanic heat transport dominates, with changes of up to ~ 0.30 °C over 4 yr. The technique is subsequently applied to observational data. For the ocean heat fluxes, we use data from the RAPID program at 26° N from April 2004 to January 2011. At 36° N heat transport is inferred using a linear regression model based on the oceanic low-frequency transport in OCCAM. The air-sea flux from OCCAM is used for the period 2004 to 2006 when the RAPID timeseries and the OCCAM simulation overlap, and a climatology is used for the air-sea flux from 2006 onwards. The results confirm that on longer (> 2 yr) timescales the ocean dominates the ocean heat content variability, which is further verified using data from the ARGO project. This work illustrates that oceanic divergence significantly impacts the ocean heat content variability on timescales relevant for applications such as seasonal hurricane forecasts.