% I've extracted the velocity fields from the inverse model,
% along with some accompanying fields.  The output is saved
% for each hydrographic section, with the file name 
%   velfields_(sectionname).mat.
% Each file contains:
%   lon, lat: longitude and latitude (horizontal coordinates)
%   p, z: pressure (dbar) and depth (m) (vertical coordinates)
%   sal, temp: salinity (psu) and in-situ temperature (deg.C)
%   ugeo, uek: geostrophic and Ekman velocity (m/s).
%  Sign convention for velocity: if you're at lat(1),lon(1)
%   and looking at lat(length(lat)),lon(length(lon)),
%   then positive is to your left.  (I had to come up with a
%   more generic convention than "positive northward" because
%   of the arbitrary orientation of the sections.)
%  Total velocity is uek+ugeo.  Each velocity is for a station
%   pair, so ugeo(1) is between the pair lon(1) and lon(2).
%  Note that the Ekman transports are actually calculated from
%   monthly climatology, not from the hydrographic section.
%   I've converted the Ekman mass transport to Ekman velocities
%   assuming that they extend to a depth of 20m - totally
%   arbitrary.
%  Area: a matrix containing the area (m^2) of each bin.
%   This is NOT just the lateral distance times diff(z), for
%   bins beneath the deepest common depth of a station pair,
%   but rather the area of the bottom triangle in that bin.
%  You can calculate the mass transport as Area.*(ugeo+uek).