you had asked about getting this grid into Arc/Info. The steps are pretty
easy except that a calculation must be made on the grid since
ArcInfo won't recognize signed integers as a datatype:

• "topo_6.2.hdr" header file:

ncols                10800
nrows              6336
nbands            1
nbits               16
byteorder       M
layout             bil
skipbytes       0
ulxmap          1.853247800e+03
ulymap          1.174032481e+07
xdim            3.706495600e+03
ydim            3.706495600e+03
bandrowbytes    21600
totalrowbytes   21600
 
• import the grid using "imagegrid topo_6.2.bil xxbil2grid" (in Arc)

• recalculate the grid to correct for the signed/unsigned integer problem (in grid):


TMP_BATH = setnull(xxbil2grid EQ 32768, con(xxbil2grid GE 32767, (xxbil2grid - 65536), xxbil2grid))
XXGRID2BIL = con(isnull(bath_dd), -32768, tmp_bath)
 

• delete the old grid with "kill xxbil2grid" (in Arc)

PROJECTION PARAMETERS
Mark Greninger (Greninger.Mark@epamail.epa.gov):

ERMapper coordinates go from 0 to 360 degrees, while Arc/Info goes from -180 to +180 degrees.
To calculate the eastings, I used the values from the ERMapper header.
Taking the number of columns (which are 360 degrees), dividing by two, and then mutiplying by the cell size (3706.4959).
This gave me a value of 20,015,076.24, which I used for the false eastings in Kirk's projection parameters.
The center of the projection is 180 degrees.

.proj file:            projection mercator
units meters
datum none
spheroid sphere
parameters
180 00 00
00 00 00
20015076.24
0.0

> From sandwell@geosat.ucsd.edu Wed Oct 21 17:31 MET 1998
>
> Kirk,
>
> Thanks for checking this.  The projection that we used
> does not depend on the radius of the earth so I don't know
> why it is needed by ER_mapper.  I don't really understand the inner workings
> of ER_Mapper.  Here is the code to go from pixel space into lat lon space
> and vice versa.  The projection has no radius or flattening.

From: kirk@geo.uni-potsdam.de (Kirk Haselton)
To: sandwell@geosat.ucsd.edu
Subject: Re: Marine topo projection and datum

Hello David,

   my understanding is that any projection you do has to rely on some model
   of the earth, whether it's a sphere (of a certain radius) or a spheroid
   like the WGS84. I'm guessing the code you use for projecting to MRWORLD
   has these coefficients hard-wired somewhere.
-------------------------------------------------------
Kirk again - I am still concerned about what datum was used to produce the original data,
the .ers file at the web site indicates "SPHERE" and yet the projection 'MRWORLD'
as defined in our version of ERMapper has a spheroidal datum (primary radius of
6,378,388 is close to 6,371,000 indicated in the Sandwell documentation at the web site).

My notes regarding the ERMapper definitions of this projection and datums:

 1) the radius for the MRWORLD projection is listed as 6378388
    while the radius for the defined datum "SPHERE" is 6371000

 2) the center latitude is 0.0, ie - equator

 3) the prime meridian index is 1 = Greenwich Merician (0.0)

mercator.dat:
proj_name,false_north,false_east,scale_factor,centre_merid,centre_lat,radius,eccentricity,flattening,prime_merid
MRWORLD ,0.0 ,20000000.0 ,1.0 ,0.0 ,0.0 ,6378388.0 ,0.0819918899790298,297.0 ,1

project.dat:
proj_name,type,metre_factor,radius,eccentricity,flattening,prime_merid
MRWORLD ,mercator,1.0 ,6378388.0 ,0.0819918899790298,297.0 ,1

projinst.dat:
proj_name,proj_type,length_id,angle_id,usearea_id,type_desc,description,ref_proj,source,visibility,release
MRWORLD ,mercator,1 ,1 ,1,01_mercator ,,,RJL AMENDED 12/9/84. FOR STORAGE WEST OF0 E; RECOMMENDED FOR ALL-WORLD RETRIEVAL,WORLD,PUBLIC
 

and yet datum is defined as sphere:
SPHERE ,0,6371000.0 ,0.0 ,1.0 ,1.0 ,1 ,0.0
( which is the same as
  USSPHERE,0,6371000.0 ,0.0 ,1.0 ,1.0 ,1 ,0.0
  and is used for reprojecting to geographic projection, or others)