% Rough diary of the steps fpr setting up a simple 
% cartesian grid for the Mullica river.

define grid lat/lon limits 
	enter in file boundary1.dat

configure coastline file
tmp = load('north_mask.txt');
latn = tmp(:,1:3)*[1 1/60 1/3600]';
lonn = tmp(:,4:6)*[1 1/60 1/3600]';

tmp = load('south_mask.txt');
lats = tmp(:,1:3)*[1 1/60 1/3600]';
lons = tmp(:,4:6)*[1 1/60 1/3600]';

lon = [lonn; NaN; lons];
lat = [latn; NaN; lats];

save mullica_coast lon lat

<seagrid>
        load coastline file
                mullica_coast.mat
click 4 times to make any old grid
<seagrid>
        load boundary points
                boundary1.dat

rubber band to zoom (if necessary)
<View>
        <Setup...>
                cells edge 1 32
                cells edge 2 64
<seagrid>
        <save as>
                seagrid.mat

grd_title = 'Mullica 1'
grd_file  = 'mullica_grid_1.nc'
seagrid2roms('seagrid',grd_file,grd_title);



Create the Land/sea mask with editmask.m

addpath /home/wilkin.new/roms/nena/mat/rmask -end

% Need to meddle with editmask.m and ijcoast.m
% There is a precision issue with the griddata triangulation
% failing for such small grid spacing. Applying an offset
% solves the problem. Tnis could be applied routinely in
% ijcoast to make the code robust
C.Icst = griddata(rlon+74,rlat-39,x,C.lon+74,C.lat-39)
C.Jcst = griddata(rlon+74,rlat-39,y,C.lon+74,C.lat-39)

interactively edit the mask, then save



bathy = load('bathymlw.txt');
lonb = bathy(:,2);
latb = bathy(:,1);
hb   = bathy(:,3);