2

I am using LAI values to assess vegetation growth over large areas. I want to be able to identify areas that are doing well and areas that are decreasing based on the changes in pixel values. Ideally I would like to generate another raster file that shows a standard deviation value over the 50 different raster layers I have, or something along those lines.

  • Just note that computing stdev is only an index of the magnitude of the variability relatively to the mean value. Yet it won't help to determine the direction, e.g. is the LAI increasing or decreasing overtime. In addition, extreme LAI index will distort your results. – dof1985 Apr 8 '15 at 11:05
  • Thanks for your comment. Can you suggest the correct method to illustrate the LAI increasing or decreasing overtime that might work with QGIS. I currently have the LAI values as band values 0-255.I have found a list of options in a grass tool for an aggregate operation - average,count, median, mode, minimum, min_raster, maximum, max_raster, stddev, range, sum, variance, diversity, slope, offset, detcoef, quart1, quart 3, perc90, skewness, and kurtosis. would any of those do the job cheers – Allan Coan Apr 9 '15 at 3:18
  • for graphical representationm I can only think of the difference raster between two selected years. You can also use the stdev to indicate about hotspots of change, either sudden and extreme, or moderate change. It would be best to extract LAI values from those areas and insepct in a line time-series graph. Those are all statistics problems. – dof1985 Apr 9 '15 at 7:35
2

You should be looking at the GRASS module r.series

Let's say you have LAI rasters in GRASS named lai_201501, lai_201412, lai_201411, etc... then at the GRASS command prompt you can run:

lai_list = `g.list type=rast pat="lai_*"`
r.series input=$lai_list output=lai_stddev method=stddev
  • Thanks a lot. Really helpful. I found the grass tool you mentioned under toolbox in the QGIS menu so no need for a command. This allowed me just to select all rasters with in my project. It seems stddev isnt what Im after though what would you suggest that would best illustrate a change in time with LAI. I currently have the LAI values indexed as band values 0-255.the options in the module are average,count, median, mode, minimum, min_raster, maximum, max_raster, stddev, range, sum, variance, diversity, slope, offset, detcoef, quart1, quart 3, perc90, skewness, and kurtosis. thanks – Allan Coan Apr 9 '15 at 3:20
  • i have had a play and the slope option seemed to work really well thanks – Allan Coan Apr 9 '15 at 4:42

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.