Raster Band Separation

[wyatt_robbins]

I'm using 366 band floating point raster data. Each band represents one day's climate data. I need to find a way to select around 100 of these bands (one season) in a manner that will not take years to complete, in order to view statistics for a season rather than a whole year. Is there any way of doing this and then being able to narrow this data down to the county level?

[DuncanHornby]

Sounds just like the sort of thing that can be automated in modelbuilder, what have you tried?

[DarrenWiens2]

Here's how you can combine a sample of random raster bands into a new raster using Python. This assumes your bands are named "Band_1", "Band_2", etc. Change the numbers to suit your data - I don't have a 366 band raster to play with.

>>> import os, random
... bands = []
... raster = r'C:\junk\comp_rast.tif'
... bandNumList = []
... for i in range(1,7): # create list from 1 - 6
...    bandNumList.append(i)
... print bandNumList
... random.shuffle(bandNumList) # randomize list
... print bandNumList
... for i in range(3): # loop through first three in list
...    bandx = os.path.join(raster,'Band_' + str(bandNumList)) 
...    bands.append(bandx) # add appropriate band path to list of raster band paths
... print bands
... arcpy.CompositeBands_management(bands,r'C:/junk/three.tif') # create new composite raster
...
[1, 2, 3, 4, 5, 6]
[3, 1, 2, 6, 4, 5]
['C:\\junk\\comp_rast.tif\\Band_3', 'C:\\junk\\comp_rast.tif\\Band_1', 'C:\\junk\\comp_rast.tif\\Band_2']

[wyatt_robbins]

I've tried some simple tools involving rasters. spatial analysis, raster creation, raster processing, conversion tools. I'm very unfamiliar with python. I have access to use it, but have never had to do so at this point, so I'm more looking for a tool of some sort.

[DarrenWiens2]

My example was overly complicated, selecting random bands and combining into one raster, but I don't think you're going to find a tool that does what you want without some customization. Below is code that will calculate the mean value for a range of bands and output that raster.

Python doesn't have to be scary. Just open the Python window in ArcMap and copy/paste. Change the band numbers and raster paths to suit your own needs (just the values after '=' in lines 2 - 5). Text after '#' is a Python comment and will be ignored.

>>> import os # import a library for later
... startBand = 1 # starting band number
... endBand = 5 # end band number
... inputRaster = r'C:\junk\YourInputRaster.tif' # input multiband raster
... outputRaster = r'C:\junk\YourOutputRaster.tif' # output single band raster
... rasterAdder = arcpy.Raster(os.path.join(inputRaster ,'Band_' + str(startBand)))
... for i in range(startBand+1,endBand+1): # loop through bands
...    bandx = arcpy.Raster(os.path.join(inputRaster ,'Band_' + str(i))) # add bands together
...    rasterAdder += bandx
... meanRaster = rasterAdder/((endBand - startBand) + 1) # calculate mean
... meanRaster.save(outputRaster) # save raster