[UW-GIS-L] Re: floating point raster map algebra / raster calculator problem

[Andy Ritchie]

I've done plenty of floating point math on 32 bit fp rasters. It sounds like
that's what you're working with since your NoData value = -3.4028235e+038. I
don't think that's the problem.

I have had weird errors when trying to do raster calculations when either
the projection or the bit depth/data types are different, or if rasters are
in different projections than each other or the data frame.

You might want to also check your raster analysis and environment settings
and see if there's anything wonky there.

Andy




On Tue, Mar 30, 2010 at 12:01 PM, <
uw-gis-l-request at mailman1.u.washington.edu> wrote:

>

> ------------------------------

>

> Message: 2

> Date: Tue, 30 Mar 2010 09:44:49 -0700

> From: Jeremy Littell <jlittell at u.washington.edu>

> Subject: [UW-GIS-L] floating point raster map algebra / raster

>        calculator      problem

> To: uw-gis-l at u.washington.edu

> Message-ID: <E74EDCBF-55DB-4F3A-B544-54D7D45B40FE at u.washington.edu>

> Content-Type: text/plain; charset="us-ascii"

>

> I've searched for help on this in the archives and didn't find

> anything, but that may be because I'm not characterizing the problem

> accurately. Any help would be appreciated.

>

> I'm running ArcMap 9.3 on Windows XP inside a virtual machine (Fusion

> on a Mac OSX 10.5). It has worked fine in the past.

>

> The problem goes like this:

>

> I import several ascii grids to floating point rasters, either command

> line or using the conversion toolbox. They are the same domain, same

> grid - all that varies is the data values. They import and display

> fine and the data values are all plausible, although in the properties

> tab under "source", I see that the NoData value = -3.4028235e+038 when

> I thought it should be -9999.

>

> When I try to do any operations on multiple grids, e.g. [grid1] -

> [grid2] , I end up with a raster that has 0 and 1 values even though

> the original data is numeric, floating point, from 0 to 2400.9 or so.

> That NoData value seems like the culprit, but I need to know where

> it's coming from so I can fix it. I want to automate this because I've

> got several hundred such grids to convert and do operations on.

>

> Thanks for any insight.

>

> JL

>

>

> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~

> Jeremy S. Littell

> Research Scientist

> CSES Climate Impacts Group

> Box 355672

> University of Washington

> Seattle, WA 98195-5672

> P: 206 221 2997

> Email: jlittell at u.washington.edu

> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~

>

> -------------- next part --------------

> An HTML attachment was scrubbed...

> URL:

> http://mailman1.u.washington.edu/pipermail/uw-gis-l/attachments/20100330/f7143e3a/attachment-0001.html

>

> ------------------------------

>

> Message: 3

> Date: Tue, 30 Mar 2010 10:14:26 -0700

> From: Michalis Avraam <michalis.avraam at gmail.com>

> Subject: Re: [UW-GIS-L] floating point raster map algebra / raster

>        calculator      problem

> To: "UW - GIS Discussion &amp, Support" <uw-gis-l at u.washington.edu>

> Message-ID:

>        <829a3c191003301014v2d47e9dpe20f6248b83f1733 at mail.gmail.com>

> Content-Type: text/plain; charset=UTF-8

>

> I have seen multiple threads on the ESRI website about floating point

> raster file errors (division seems to return integer values too for

> some reason).

>

> The solution most people suggest is to transform your floating point

> rasters to integer rasters (as it also save a bunch of space due to

> compression), and then perform your calculations. If you only have one

> decimal place, then just multiply everything by 10 to get rid of the

> decimal, do your operation, and if you need to, divide by 10 again

> (10.0 probably and cast as float() to make it work).

>

> For your null values, I suggest you use the SetNull() command

> (http://webhelp.esri.com/arcgisdesktop/9.3/index.cfm?TopicName=Set_Null)

> to get the proper Null Value you may need. This could solve your

> original problem, but from what I see in the forums, it may not do a

> difference.

>

> Another thing you may consider is the domain of your data. NoData in

> the ESRI world is actually a data value set as NoData. If your data

> cover the full floating point 32-bit domain (unlikely, but you never

> know), then ArcGIS cannot represent NoData (until the new 64-bit data

> appear). As an aside, if you use 1bit raster and decide to create Null

> values, then the raster gets stored as 2bit, and similarly it moves on

> until floating point 32bit and then stops.

>

> Hope this helps and wasn't more confusing.

>

> Michalis

>

>

>

>

>

> 2010/3/30 Jeremy Littell <jlittell at u.washington.edu>:

> > I've searched for help on this in the archives and didn't find anything,

> but

> > that may be because I'm not characterizing the problem accurately. Any

> help

> > would be appreciated.

> > I'm running ArcMap 9.3 on Windows XP inside a virtual machine (Fusion on

> a

> > Mac OSX 10.5). It has worked fine in the past.

> > The problem goes like this:

> > I import several ascii grids to floating point rasters, either command

> line

> > or using the conversion toolbox. They are the same domain, same grid -

> all

> > that varies is the data values. They import and display fine and the data

> > values are all plausible, although in the properties tab under "source",

> I

> > see that the NoData value = -3.4028235e+038 when I thought it should be

> > -9999.

> > When I try to do any operations on multiple grids, e.g. [grid1] - [grid2]

> ,

> > I end up with a raster that has 0 and 1 values even though the original

> data

> > is numeric, floating point, from 0 to 2400.9 or so. That NoData value

> seems

> > like the culprit, but I need to know where it's coming from so I can fix

> it.

> > I want to automate this because I've got several hundred such grids to

> > convert and do operations on.

> > Thanks for any insight.

> > JL

> >

> > ~~~~~~~~~~~~~~~~~~~~~~~~~~~~

> > Jeremy S. Littell

> > Research Scientist

> > CSES Climate Impacts Group

> > Box 355672

> > University of Washington

> > Seattle, WA 98195-5672

> > P: 206 221 2997

> > Email: jlittell at u.washington.edu

> > ~~~~~~~~~~~~~~~~~~~~~~~~~~~~

> >

> > _______________________________________________

> > Uw-gis-l mailing list

> > Uw-gis-l at u.washington.edu

> > http://mailman1.u.washington.edu/mailman/listinfo/uw-gis-l

> >

> >

>

>

> ------------------------------

>

> _______________________________________________

> Uw-gis-l mailing list

> Uw-gis-l at mailman1.u.washington.edu

> http://mailman1.u.washington.edu/mailman/listinfo/uw-gis-l

>

>

> End of Uw-gis-l Digest, Vol 92, Issue 13

> ****************************************

>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mailman1.u.washington.edu/pipermail/uw-gis-l/attachments/20100331/6c949efd/attachment.html