NLCD2006 Raster file skewed

283
5
07-11-2012 10:07 AM
Highlighted
New Contributor
Hello,

I'm having an issue with a raster file pulled from seamless.usgs.gov. I attached an image showing my problem. The NLCD raster is skewed compared to my area of study (blue area). The raster has a Spatial Reference: USA Contiguous Albers Equal Area Conic USGS Version , whereas the area of study has a projection NAD_1983_HARN_StatePlane_Illinois_East_FIPS_1201. I have tried changing the raster spatial reference, which "corrects" the skewedness however places it in a very different location in ArcMap.

I've tried a bunch of different options from Rotating the Raster to trying different ways of clipping the raster data to align with my area of study. Any thoughts would be appreciated! I'm still learning all of this so i usually don't post a question unless i hit a wall....and I have been hitting the wall for the past two days.

Thank you in advance!

pat
Reply
0 Kudos
5 Replies
Highlighted
Regular Contributor
What transformation are you using in the map document?  Have you considered projecting the data into Illinois Harn?  I would recommend either of those options, depending on your workflow.  Setting the correct transformation in the map document or using the correct transformation to project the data should fix the "skew."  In this blog, it discusses how to choose the desired transformation.
Reply
0 Kudos
Highlighted
New Contributor
Hi,Patrick

I suggest you import the NLCD's projection into the transfer process and make sure
choose the "Bilinear" resampling method.
I believe it will works!

Xi
Reply
0 Kudos
Highlighted
New Contributor
Thank you for your responses. I'm still missing something though. I have tried setting the Data Frame Coordinate System to the IL_HARN, then adding the raster data through transforming the NLCD raster file -> Convert From: GCS_North_American_1983   Into: GCS_North_American_1983_HARN  Using:  NAD_1983_to_HARN_Illlinois

Still it doesn't work correctly, if the NLCD is not skewed, it is placed in a completely different area than the rest of my data. One thing I noticed with the NLCD data is TIFF with world file. When i look at the projected coordinate system properties, it is using the "Albers" Projection and the Geographic Coordinate System: GCS_North_American_1983 Datum: D_North_American_1983.

I attempted to change this file to the HARN_Illinois AND the raster file Spatial Reference to HARN_Illinois. When I did this it placed the raster data in a completely different place but the data wasn't skewed.

I honestly don't know what I am missing. I have resampled the data using BILINEAR after importing with transformation but it still isn't giving me what i need....

Any thoughts?
Reply
0 Kudos
Highlighted
Regular Contributor
For specific help then if those suggestions do not work, then I recommend creating a Support Incident to look over your workflow. Something is still not correct, because while the skewing should happen without the transformation, you should see some effect from either projecting the raster or using the proper transformation.
Reply
0 Kudos
Highlighted
New Contributor
thanks for your help. I always feel there are 10,000 different ways of doing something in ArcGIS. I found a thread (http://forums.arcgis.com/threads/25476-Project-USA-Contiguous-Albers-Equal-Area-Conic-USGS-into-Lamb...) that suggested the following: "Set the data frame's coordinate system to the custom Lambert azimuthal equal area. Ignore the warning message about datums are different. Right click the layer and use data, export data and save the raster with the data frame's coordinate system. "

I did those same steps but for my coordinate system and got some decent results but still not perfect. I posted another picture of what I have now. I remember seeing in a different thread (when i was searching for my initial problem), someone discussing issues with Extract by Mask and Snap Raster. So I'm going to look up those to see if I can't correct my "new" problem.

Thank you again. and any suggestions you might have on my new problem would be appreciated.

Pat
Reply
0 Kudos