Site Selection Analysis Workflow

Question

What would be an appropriate workflow for a site selection analysis when beginning with rasters of mixed cell size and different units of measurement (feet and meters)?  It's been a long time since I've done one of these and I know the basic workflow when starting with matching data sets but before that I'm unsure.  Any help would be appreciated.

DanPatterson_Retired · Answer

consistency and documentation is the key Annie.  Anything that you skip or gloss over may have to be redone later... so choosing your cell size, extent and those other issues first will lay the groundwork for subsequent processing and analysis... good luck

AnnieCahill · Answer

Thanks for your reply.  The only real file I have right now for this project is a land cover raster at 1 meter resolution.  The other data sets (DEM and slope) I was planning on creating from our 2 foot contour data, but I was unsure if I should match the cell sizes of the new rasters I create to the land cover (it sounds like I should).  Our coordinate system is State Plane (feet).  Should change my unit of measurement to feet on the land cover data set and then generate a DEM from my contours to match the cell size of the land cover?  I think I can set up most of this in the environment settings if I remember correctly.  Thanks again so much for your help!

DanPatterson_Retired · Answer

some steps can be combined, but I would

* convert rasters to the desired system of units (probably meters) while
* resampling the coarser rasters to match the finer rasters while
* ensuring that you specify a common extend while
* specifying a snap raster to ensure cell alignments

You didn't mention different projections... that is a tough call, but I would project to a common coordinate system in the desired planar units.  There may be issues with maintaining a couple of the point above consistently.

In any event,

* always set and check your raster analysis and extent properties in the Environments tab within every tool that you use. 
* Check the help topic for any tool you use at the very bottom and there is usually some specification as to the environments that are supported by that tool.
* Take your worst file, run the process and compare it to one that needs no work and if there are common spatial elements compare what happened to them during the process.

I probably have missed some and/or got a few out of order... again a few can be combined into one step.... if you aren't sure, do the steps individually.  and... get coffee and make notes on paper

Addendum

The obvious... short paths, local data sources and destinations, no spaces, pay attention to dat type (ie avoid bit upscaling when possible, try a common nodata value when possible),  In short... anything that will foul a result