r.resample

[Julie Daniel]

Hello GRASS users,

I have done some experiments about what happens to raster maps when you
change the resolution of your region and then run r.resample.
I work in a Lat/Lon projection and normally use a resolution of 15 seconds.
In a 1x1 minute area, there are sixteen 15 second raster cells.
I changed my resolution to 30 seconds (so there are now four 30 second raster c
ells)
then ran r.resample on my map.

This is what my map looked like at a resolution of 15 seconds:

                     |----|----|----|----|41:01N
                     | 1  | 2  | 3  |  4 |
                     |----|----|----|----|41:00:45N
                     | 5  | 6  | 7  |  8 |
                     |----|----|----|----|41:00:30N
                     | 9  | 10 | 11 | 12 |
                     |----|----|----|----|41:00:15N
                     | 13 | 14 | 15 | 16 |
                     |----|----|----|----|41:00N

This is what I thought my result would be at 30 seconds:
(take the value of the upper-left cell out of the center four)

                     |---------|---------|41:01N
                     |         |         |
                     |    1    |    3    |
                     |         |         |
                     |---------|---------|41:00:30N
                     |         |         |
                     |    9    |   11    |
                     |         |         |
                     |---------|---------|41:00N

But, what I got was this:

                     |---------|---------|
                     |         |         |
                     |    6    |    8    |
                     |         |         |
                     |---------|---------|
                     |         |         |
                     |   14    |   16    |
                     |         |         |
                     |---------|---------|


Can anyone explain why GRASS used the bottom-right cell out of the center four
and not the upper-left cell?
I can't find any reference of what GRASS does an any manual, but I've read on t
his
user group that it takes the upper-left cell.
Any help would be sincerely appreciated.
Thanks so much.

Julie Daniel
Physical Scientist
National Weather Service

(301) 713-1669
email: julied at lionheart.ssmc.noaa.gov