G S H H S
Global Self-consistent Hierarchical High-resolution Shorelines
Distributed under the GNU Public License (see file ../LICENSE.TXT)
Version Jan 1, 2011
No source code changes but expected to work with the latest GSHHS
data set version 2.1.1
Version July 1, 2010
No source code changes but expected to work with the latest GSHHS
data set version 2.1.0
Version July 1, 2009
This version include a revised GSHHS data format in which the
polygon ID of the parent (that contains this polygon) is given
(or -1 if none), as well as the ancestor (Id of the full resolution
polygon that yielded this reduced polygon), and the area of the
reduced resolution polygon. We also keep track if a lake is a
river-lake or a regular lake. The header structure has thus
increased by three integer items. The GSHHS data set is now
version 2.0; from now on we will no longer give version number for
the gshhs supplement programs since it leads to confusion with the
GSHHS data version numbers. See gshhs.h for new structure.
Version 1.11 Feb 15, 2009
This version adds the -I<id> option go gshhs so that a single
polygon can be extracted.
Also, thanx to Markus Metz,GSHHS shorelines can now be imported
directly into GRASS with v.in.gshhs, a GRASS AddOn available on
http://grass.osgeo.org/wiki/GRASS_AddOns. This module
works with GRASS 6.4 and GSHHS databases version 1.4 and above.
As an alternative there is still gshhstograss, a standalone program
creating a simple GRASS ASCII vector. It is recommended to use
v.in.gshhs instead.
Version 1.10 May 1, 2008
Due to a numbering snafu the GMT 4.2.1 release started calling
the GSHHS data set using the same numbering as the GSHHS software.
Thus, GSHHS v 1.9 is the data set version that follows v1.6.
For the GMT 4.3.0 release the GSHHS version is 1.10 since we
have added two missing islands off Long Island, NY.
Also, thanx to Markus Metz the gshhstograss.c has been updated.
--------------------------------------------------------------
Version 1.9 Sept 1, 2007
Now distributing the WDB borders and rivers with GSHHS, hence
gshhs.c needed a few tweaks to recognize the difference. No
data format change, but we increment version number to 1.6 do
indicate a change in composition (river+borders added).
--------------------------------------------------------------
Version 1.8 March 31, 2007
Minor fix to ensure gshhstograss.c will compile. Version 1.5
of the GSHHS data released, fixing a few crossings and dupli-
cate points. No data format change so 1.7 can read as well.
--------------------------------------------------------------
Version 1.7 Nov 11, 2006
Minor bug fix: gshhs.c used && instead of & in one of the bit
manipulations, thus always yielding 1 as the level.
--------------------------------------------------------------
Version 1.6 May 13, 2006
Minor update to handle the latest GSHHS v1.4 database. Once again
we claim that now, for sure, GSHHS is finally self-consistent!
Due to a bug in our consistency checker a few lingering problems
made it into v1.3 - these should now all be gone. Many thanx to
Thomas DeMay for contributing the results from his error checking.
We have also once again made changes to the GSHHS header structure
to make it only contain integer*4 (and not a mix of 2- and 4-byte
integers) as well as be a multiple of 16 bytes. Thus, the new
structure is
int id; /* Unique polygon id number, starting at 0 */
int n; /* Number of points in this polygon */
int flag; /* level + version << 8 + greenwich << 16 + source << 24
int west, east, south, north; /* min/max extent in micro-degrees */
int area; /* Area of polygon in 1/10 km^2 */
Here, level, version, greenwhich, and source are
level: 1 land, 2 lake, 3 island_in_lake, 4 pond_in_island_in_lake
version: Set to 4 for GSHHS version 1.4
greenwich: 1 if Greenwich is crossed
source: 0 = CIA WDBII, 1 = WVS
These values are recovered using bit math. For further info about
GSHHS see the earlier comments below.
--------------------------------------------------------------
Version 1.5 Sept 14, 2004
Minor update to handle the latest GSHHS v1.3 database. GSHHS is now
[finally] self-consistent: There are no longer any lingering cross-
overs (internal or external), all polygons are explicitly closed,
and there are no duplicate points along the lines. The GSHHS header
record has been expanded by one 4-byte element (called version and
set to 3 for all polygons) so that the size of the entire header
structure is a multiple of 8 bytes (now 40 bytes). The old structure
was not as portable because padding of structures is architecture-
dependent. To be explicit, each header in v1.3 now consist of the
following variables:
int id; /* Unique polygon id number, starting at 0 */
int n; /* Number of points in this polygon */
int level; /* 1 land, 2 lake, 3 island_in_lake, 4 pond_in_island_in_lake */
int west, east, south, north; /* min/max extent in micro-degrees */
int area; /* Area of polygon in 1/10 km^2 */
int version; /* Polygon version, set to 3
short int greenwich; /* Greenwich is 1 if Greenwich is crossed */
short int source; /* 0 = CIA WDBII, 1 = WVS */
We remind users that the polygons are sorted by polygon length (not
area), and that the w/e/s/n values for each polygon are those based on
the full resolution and are simply copied to the other resolutions.
The decimated polygons will in general have a different w/e/s/n set
which you can find if you calculate the actual w/e/s/n values.
Allthough gshhs is distributed as a GMT supplement, it makes no references
to GMT include files or libraries. If you need to compile gshhs without
placing the gshhs directory under GMT/src you can run
cc gshhs.c -O2 -lm -o gshhs
cc gshhs_dp.c -O2 -lm -o gshhs_dp
--------------------------------------------------------------
Version 1.4 Sept 5, 2000
Now GSHHS source code is distributed as a GMT supplement. The data
may be obtained separately from the SOEST or NGDC web sites as before.
The FLIP macro is no longer needed as the programs will determine if
swab'ing is required. Do no use dd on the data.
--------------------------------------------------------------
Version 1.3 Nov 8, 1999
Now all code and data files are explicitly distributed under the
GNU Public License, see file LICENSE.TXT for more details.
--------------------------------------------------------------
Version 1.2 May 18, 1999
Made programs POSIX.1 compliant and added binary open for DOS.
--------------------------------------------------------------
Version 1.1, April 30, 1996
Paul Wessel, G&G, SOEST, U of Hawaii (pwessel@hawaii.edu)
Walter H. F. Smith, NOAA Geosciences Lab (walter@raptor.grdl.noaa.gov)
Ref: Wessel, P., and W. H. F. Smith, 1996, A global self-consistent,
hierarchical, high-resolution shoreline database, J. Geophys.
Res., 101, 8741-8743.
For details on data processing etc. we refer you to that reference.
--------------------------------------------------------------------
This README file explains the usage of the gshhs data sets. The
archive consists of the following files (after you unzip the compressed
files using bzip2 -d):
Name Content
--------------------------------------------------------------------
README This file
gshhs.h Header file for programs
gshhs.c Program to extract ASCII data
gshhs_dp.c Program to decimate polygons
gshhs_f.b Full resolution data
gshhs_h.b High resolution data
gshhs_i.b Intermediate resolution data
gshhs_l.b Low resolution data
gshhs_c.b Crude resolution data
In addition, the following program was supplied by Simon Cox (simon@ned.dem.csiro.au)
and can be used to import the *.b files into a GRASS database:
gshhstograss.c Import *.b into GRASS GIS database
All the *.b file share the same file structure; thus the gshhs program can
read and extract data from any of the files. The program's purpose is
simply to demonstrate how a programmer may access the data. Presumably,
the user wants to access the data from within his/her own programs.
If plotting the data is the only purpose, we strongly recommend you
instead use the GMT package which comes with the same data and tools for
plotting filled landmasses, coastlines, political borders, and rivers.
The file(s) contain several successive logical blocks of the form
<polygon header>
<polygon points>
Each header consist of the following variables:
int id; /* Unique polygon id number, starting at 0 */
int n; /* Number of points in this polygon */
int level; /* 1 land, 2 lake, 3 island_in_lake, 4 pond_in_island_in_lake */
int west, east, south, north; /* min/max extent in micro-degrees */
int area; /* Area of polygon in 1/10 km^2 */
short int greenwich; /* Greenwich is 1 if Greenwich is crossed */
short int source; /* 0 = CIA WDBII, 1 = WVS */
Here, int is 4-byte integers and short means 2-byte integers.
The polygon points are stored as n successive records of the form
int x; /* longitude of a point in micro-degrees */
int y; /* latitude of a point in micro-degrees */
On some systems, the byte order is swapped relative to the order used on
a Sun workstation (on which the current data were processed). To
determine if you need to swap the byte pairs, do the following test:
1. Compile gshhs
2. Run gshhs gshhs_c.b | head -1 # This shows the 1st line of output
3. If the output does not look exactly like the next line:
P 0 1240 1 W 79793839.900 -17.53378 190.32600 -34.83044 77.71625
you most likely need to swap the byte-pairs. Simply recompile gsggs with
the switch -DFLIP and see if that did the trick.
4. If all fails you may email one of the authors for advice.
Compile the two programs as follows (with or without the -DFLIP switch):
cc -O gshhs.c -o gshhs [-DFLIP]
cc -O gshhs_dp.c -o gshhs_dp -lm [-DFLIP]
[and optionally cc -O -o gshhstograss gshhstograss.c -lm [-DFLIP]]
We have provided 5 different resolution of the data which should
satisfy just about any user. The [h,i,l,c]-versions were
derived from the gshhs_f.b full resolution file using the
Douglas-Peucker algorithm as implemented in gshhs_dp.c The
tolerances used were:
File Content Tolerance
-------------------------------------------------
gshhs_h.c High resolution 0.2 km
gshhs_i.c Interm. resolution 1.0 km
gshhs_l.c Low resolution 5.0 km
gshhs_c.c Crude resolution 25 km
However, should you need to decimate the full data set using a
different tolerance you can use the program gshhs_dp to do so:
gshhs_dp gshhs_f.b your_tolerance_in_km newfile.b
gshhs.c can then read the resulting newfile.b
[Note that output from gshhs_dp WILL NOT need byte swapping since
it is created on your machine].
The Douglas-Peucker routine implemented in gshhs_dp was kindly
provided by Dr. Gary J. Robinson, Environmental Systems Science Centre,
University of Reading, Reading, UK (gazza@mail.nerc-nutis.ac.uk).
Good Luck,
Paul Wessel and Walter. H. F. Smith
GMT URL: http://gmt.soest.hawaii.edu
