Review of Global Ocean Intermediate Water Masses: 1.Part A, the Neutral Density Surface (the ‘McDougall Surface’) as a Study Frame for Water-Mass Analysis

This review article commences with a comprehensive historical review of the evolution and application of various density surfaces in atmospheric and oceanic studies. The background provides a basis for the birth of the neutral density idea. Attention is paid to the development of the neutral density surface concept from the nonlinearity of the equation of state of seawater. The definition and properties of neutral density surface are described in detail as developed from the equations of state of seawater and the buoyancy frequency when the squared buoyancy frequency N2 is zero, a neutral state of stability. In order to apply the neutral density surface to intermediate water-mass analysis, this review also describes in detail its practical oceanographic application. The mapping technique is focused for the first time on applying regularly gridded data in this review. It is reviewed how a backbone and ribs framework was designed to flesh out from a reference cast and first mapped the global neutral surfaces in the world’s oceans. Several mapped neutral density surfaces are presented as examples for each world ocean. The water-mass property is analyzed in each ocean at mid-depth. The characteristics of neutral density surfaces are compared with those of potential density surfaces.

Mapping of an Approximate Neutral Density Surface with Ungridded Data

A neutral density surface is a logical study frame for water-mass mixing since water parcels spread along such a surface without doing work against buoyancy restoring force. Mesoscale eddies are believed to stir and subsequently mix predominantly along such surfaces. Because of the nonlinear nature of the equation of state of seawater, the process of accurately mapping a neutral density surface necessarily involves lateral computation from one conductivity, temperature and depth (CTD) cast to the next in a logical sequence. By contrast, the depth of a potential density surface on any CTD cast is found solely from the data on this cast. The lateral calculation procedure causes a significant inconvenience. In a previous paper by present author published in this journal (You, 2006), the mapping of neutral density surfaces with regularly gridded data such as Levitus data has been introduced. In this note, I present a new method to find the depth of a neutral density surface from a cast without having to specify an integration path in space. An appropriate reference point is required that is on the neutral density surface and thereafter the neutral density surface can be de- termined by using the CTD casts in any order. This method is only approximate and the likely errors can be estimated by plotting a scatter diagram of all the pressures and potential temperatures on the neutral density surfaces. The method assumes that the variations of potential temperature and pressure (with respect to the values at the reference point) on the neutral density surface are proportional. It is important to select the most appropriate reference point in order to approximately satisfy this assumption, and in practice this is found by inspecting the θ-p plot of data on the surface. This may require that the algorithm be used twice. When the straight lines on the θ-p plot, drawn from the reference point to other points on the neutral density surface, enclose an area that is external to the clus- ter of θ-p points of the neutral density surface, errors will occur, and these errors can be quantified from this diagram. Examples showing the use of the method are presented for each of the world’s main oceans.