#1 Thu 09 March 2000 14:18
- Julie CHARLEUX
- Invité
ZONALCENTROID
Good morning,
I would like to have precision on ZONALCENTROID of the GRID module.
I study watersheds’ geometry using a DEM and a coverage of outlets.
In the GRID module, I have extracted the basins for each outlet with
WATERSHED, to have one grid for each basin.
Then, I have used ZONALCENTROID to obtain a new grid defining the
equivalent ellipse for each basin, its centroid and its axis.
For many basins, I have got an error message, because the new grid had
nodata everywhere, so could not be defined.
One amazing thing is that when I use the ZONALGEOMETRY with the ellipse
option, I get a statistical table with erroneous ellipse parameters, but
no error message.
To solve the problem with ZONALCENTROID, I have tried :
- to round a little the limits of the basin, with a FOCALMAX and then,
- to reduce the study area using a boundary obtained from the polygon
coverage of the same basin (determined by GRIDPOLY).
And, for some basins, it seems to work well. But, not for every basin.
I would be delighted if anybody could help me to solve this problem.
Thank you.
Best regards,
Julie Charleux
#2 Thu 09 March 2000 14:19
- Julie CHARLEUX
- Invité
Re: ZONALCENTROID
Rebonjour,
toutes mes excuses pour mon anglais dans mon precedent mail, je continue
donc en francais.
J'ai contourne le probleme de calcul de l'ellipse equivalente des
bassins en degradant les grids representant les bassins a une resolution
5 fois plus grossiere. Les parametres de l'ellipse sont quasi inchanges
et surtout le calcul fonctionne !
En effet, la zone de travail est tres etendue : 9696*8045 cellules.
Pour les petits bassins, la zone de calcul du ZONALCENTROID etait
suffisamment reduite pour obtenir une ellipse correcte.
Il semble que, a partir d'un bassin de plus de 3052000 cellules, le
calcul de l'ellipse equivalente peut planter...
La degradation des bassins permet de contourner ce probleme.
Merci a ceux qui m'ont repondu.
Julie Charleux