The Role of Advection in Determining the Temperature Structure of the Irish Sea

Abstract

The temperature structure of the Irish Sea is investigated using a 3-yr simulation with a high-resolution (?1.8 km) three-dimensional baroclinic model (the Proudman Oceanographic Laboratory Coastal-Ocean Modelling System) and CTD and Advanced Very High Resolution Radiometer observations. This paper focuses on the extent to which (horizontal) advection determines the temperature structure. It is found that it has a significant effect on the depth-mean temperatures throughout the region and on the vertical profiles in seasonally stratified areas, such as the Celtic Sea and western Irish Sea. There is depth-mean advective heating during the summer in these stratified regions, whereas in well-mixed regions advection tends to reduce the amplitude of the seasonal cycle. Through an analysis of the terms in the temperature equation, the warming of the ?cool pool? waters of the western Irish Sea can be attributed to the advection of partially well-mixed waters into the stratified region from the north. This occurs as an entrainment process with the southward current on the western side of this region folding in this water from the north. This current is seen to originate both as part of the ?gyre? circulation and from southward flow through the North Channel of the Irish Sea. The accuracy to which temperatures are modeled (particularly near the seabed in this stratified region), as compared with an experiment without temperature and salinity advection, lends weight to this interpretation of the model results. Overall rms errors against CTD observations are 1.1°C with advection and 1.7°C without. In addition to the direct effects of currents, salinity stratification (which is not present without advection in the western Irish Sea in this model) is seen to play a role in determining the temperature structure, particularly in the spring and early summer. Unlike previous baroclinic simulations in this region, the model run is continued for a further 2 yr, allowing the investigation of the seasonal cycle of temperature far removed from the initial condition. In a number of regions, a systematic overestimation of the winter temperatures is found (the cause of which has yet to be identified), but this bias does not compromise the accuracy of the results between the spring and autumn of subsequent years.