Dear GRASS users,
My problem concerns a raster in lat-lon projection, having a given resolution of ewres1,nsres1. For every cell, the category value is expressed in kilogram.
I need to transform it into the same grid but with coarser resolution, say ewres2=10*ewres1 and nsres2=10*nsres1, BUT the transformation need to be mass-conservative (so, I need to get exactly the same amount in kilograms with finer and coarser resolution).
Can I do in a simple way with GRASS ?
Thanks in advance,
Bruno Guillaume.
Bruno, first thing, in a latlong projection, the units of resolution are given in degrees. So ewres=1 means that your cell width is one degree. Most other projections have a cell unit of meters. For example an ewres=1 in an UTM projection would mean that your cell width is one meter. To change your resolution, you can use
g.region res=your_res
for example
g.region res=00:00:30 yields a 30 arc second resolution in a latlong projection.
Hope this helps,
Ian
On Oct 10, 2005, at 7:35 AM, Bruno GUILLAUME wrote:
Dear GRASS users,
My problem concerns a raster in lat-lon projection, having a given resolution of ewres1,nsres1. For every cell, the category value is expressed in kilogram.
I need to transform it into the same grid but with coarser resolution, say ewres2=10*ewres1 and nsres2=10*nsres1, BUT the transformation need to be mass-conservative (so, I need to get exactly the same amount in kilograms with finer and coarser resolution).
Can I do in a simple way with GRASS ?
Thanks in advance,
Bruno Guillaume.
>
What happens if a big asteroid hits Earth? Judging from realistic simulations involving a sledge hammer and a common laboratory frog, we can assume it will be pretty bad.
- Dave Barry
-------------------------------------------------------------
This message has been scanned by Postini anti-virus software.
Ian,
thanks for your prompt reply. Let us talk about my raster, say "toto" (unit=kg).
Well, in fact, my initial resolution for toto is 00:00:30 but it is full world map (=933120000 grid cells in total!), so it is not easily portable in flat ASCII format!! So, I prefer to convert it first to a resolution of 1°x1° (this resolution is largely enough for my goals).
The problem is that after using the g.region command to convert it, I check out the conservation of mass and my problem is that I do not get the same total amount of kg (unit of toto) with r.sum rast=toto under 00:00:30 resolution and under 1° resolution !
So, I suppose that g.region do not accumulate the field of values represented in toto (unit=kg) but rather interpolate it.
So what does it do exactly ? And how can I get it to change resolution with mass-conservation of the field in toto, so that the value in one big grid cell of toto in 1°x1° resolution be equal to the sum of all little grid cell values (of resolution 00:00:30) comprised in the big cell ?
Hope this problem is clearer expressed,
Bruno.
Ian MacMillan wrote:
Bruno, first thing, in a latlong projection, the units of resolution are given in degrees. So ewres=1 means that your cell width is one degree. Most other projections have a cell unit of meters. For example an ewres=1 in an UTM projection would mean that your cell width is one meter. To change your resolution, you can use
g.region res=your_res
for example
g.region res=00:00:30 yields a 30 arc second resolution in a latlong projection.
Hope this helps,
Ian
On Oct 10, 2005, at 7:35 AM, Bruno GUILLAUME wrote:
Dear GRASS users,
My problem concerns a raster in lat-lon projection, having a given resolution of ewres1,nsres1. For every cell, the category value is expressed in kilogram.
I need to transform it into the same grid but with coarser resolution, say ewres2=10*ewres1 and nsres2=10*nsres1, BUT the transformation need to be mass-conservative (so, I need to get exactly the same amount in kilograms with finer and coarser resolution).
Can I do in a simple way with GRASS ?
Thanks in advance,
Bruno Guillaume.
>
What happens if a big asteroid hits Earth? Judging from realistic simulations involving a sledge hammer and a common laboratory frog, we can assume it will be pretty bad.
- Dave Barry
-------------------------------------------------------------
This message has been scanned by Postini anti-virus software.
Bruno, sorry that I misidentified your problem. I think you are looking to use a command like r.neighbors. I haven't used it much, but I think there is a sum function in there that you can use.
G'luck,
Ian
On Oct 10, 2005, at 8:20 AM, Bruno GUILLAUME wrote:
Ian,
thanks for your prompt reply. Let us talk about my raster, say "toto" (unit=kg).
Well, in fact, my initial resolution for toto is 00:00:30 but it is full world map (=933120000 grid cells in total!), so it is not easily portable in flat ASCII format!! So, I prefer to convert it first to a resolution of 1°x1° (this resolution is largely enough for my goals).
The problem is that after using the g.region command to convert it, I check out the conservation of mass and my problem is that I do not get the same total amount of kg (unit of toto) with r.sum rast=toto under 00:00:30 resolution and under 1° resolution !
So, I suppose that g.region do not accumulate the field of values represented in toto (unit=kg) but rather interpolate it.
So what does it do exactly ? And how can I get it to change resolution with mass-conservation of the field in toto, so that the value in one big grid cell of toto in 1°x1° resolution be equal to the sum of all little grid cell values (of resolution 00:00:30) comprised in the big cell ?
Hope this problem is clearer expressed,
Bruno.
Ian MacMillan wrote:
Bruno, first thing, in a latlong projection, the units of resolution are given in degrees. So ewres=1 means that your cell width is one degree. Most other projections have a cell unit of meters. For example an ewres=1 in an UTM projection would mean that your cell width is one meter. To change your resolution, you can use
g.region res=your_res
for example
g.region res=00:00:30 yields a 30 arc second resolution in a latlong projection.
Hope this helps,
Ian
On Oct 10, 2005, at 7:35 AM, Bruno GUILLAUME wrote:
Dear GRASS users,
My problem concerns a raster in lat-lon projection, having a given resolution of ewres1,nsres1. For every cell, the category value is expressed in kilogram.
I need to transform it into the same grid but with coarser resolution, say ewres2=10*ewres1 and nsres2=10*nsres1, BUT the transformation need to be mass-conservative (so, I need to get exactly the same amount in kilograms with finer and coarser resolution).
Can I do in a simple way with GRASS ?
Thanks in advance,
Bruno Guillaume.
>
What happens if a big asteroid hits Earth? Judging from realistic simulations involving a sledge hammer and a common laboratory frog, we can assume it will be pretty bad.
- Dave Barry
-------------------------------------------------------------
This message has been scanned by Postini anti-virus software.
>
What happens if a big asteroid hits Earth? Judging from realistic simulations involving a sledge hammer and a common laboratory frog, we can assume it will be pretty bad.
- Dave Barry
-------------------------------------------------------------
This message has been scanned by Postini anti-virus software.
On Tue, 11 Oct 2005 01:35, Bruno GUILLAUME wrote:
Dear GRASS users,
My problem concerns a raster in lat-lon projection, having a given
resolution of ewres1,nsres1. For every cell, the category value is
expressed in kilogram.
I need to transform it into the same grid but with coarser resolution,
say ewres2=10*ewres1 and nsres2=10*nsres1, BUT the transformation need
to be mass-conservative (so, I need to get exactly the same amount in
kilograms with finer and coarser resolution).
Can I do in a simple way with GRASS ?
Depends on which version you are using.
In version 5:
r.to.sites
g.region
s.cellstats
s.to.rast
Version 6 doesn't have an equivalent to s.cellstats
Regards
Gordon
--
Gordon Keith
Programmer/Data Analyst
Marine Acoustics
CSIRO Marine and Atmospheric Research
http://www.cmar.csiro.au
42.7% of all statistics are made up.