hexagonal grid

Michael Shapiro shapiro at zorro.cecer.army.mil
Mon Nov 2 20:15:14 EST 1992


GRASS won't support hexagonal grids as it stands now. The access methods,
and all the tools have assumed (either explicitly or implicitly) rectangles.
But there is no reason why you can't use GRASS to extract a hexagonal grid
from the rectangular grid, run the model on the extracted map, and then
import any maps created by the model into grass. You would have to write
the import/export modules. The issue is whether you can (1) live with
the resampling of the data implied by the conversion process - since
with either rectangles or hexagons there is the assumption of homegenous
information in one cell, and (2) live with the import/export export process.

|
|> 
|> What is the interest in hexagonal grids? GRASS doesn't use them.
|> 
|> Of course if you had a system that used hexagonal grids you could use
|> a simple resampling to shift every other row one half cell to the right
|> (or left) and that would create a hexagonal grid.
|> 
|
|I'm trying to tie a spatially-explicit model of insect population flux, 
|movement, etc. to a real landscape via GIS.  Specifically, I'm modeling
|southern pine beetle infestations in the Georgia piedmont forests.  Hexagonal
|grids have several advantages in any model of dispersing animal populations:
|
|1)  they permit six directions of movement rather than four,
|
|2)  they fill the space completely, unlike octagons, etc, and,
|
|3)  the center to center distance to any adjacent cell is the same, regardless
|of direction, unlike rectangular grids where the diagonal distance is greater 
|than the vertical or the horizontal distances.
|
|One of our local GIS labs has worked out tools for converting ARC/INFO map
|coverages into a hexagonal array, but I prefer to work with GRASS.  Any help
|would be greatly appreciated.
|
|Note that I'm not talking about just imposing a hexagonal vector pattern onto
|raster images, but reclassifying the underlying raster map into an array of
|hexagonal cells via some sort of transformation within each hexagon, 
|e.g. "every pixel within the hexagon is set to the dominant class ocurring
|within that hexagon".  I need to apply my transformations across the entire
|map layer, hexagon by hexagon.  The biological model will then run on the
|resulting "landscape" of hexagonal cells.  Actually, it will run on a land-
|scape composed of many layers of such cells, representing the model parameters
|distributed in space.
|
|The capacity to produce these hexagonally transformed landscapes is very
|important for modeling biological movement or dispersion.
|
|
|	Michael Camann				camann at athena.cs.uga.edu
|	Department of Entomology		office 	(706) 542-2640
|	University of Georgia				(706) 542-2816
|	Athens, GA  30602			fax	(706) 542-2272
|
|


-----------------------------------------------------------------
Michael Shapiro                        U.S. Army CERL                  
email:   shapiro at amber.cecer.army.mil  Office of GRASS Integration
phone:   (217) 373-7277                P.O. Box 9005                   
fax:     (217) 373-7222                Champaign, Ill. 61826-9005
-----------------------------------------------------------------




More information about the grass-user mailing list