Take the 2-minute tour ×
Geographic Information Systems Stack Exchange is a question and answer site for cartographers, geographers and GIS professionals. It's 100% free, no registration required.

I calculated zonal statistic in Arcgis to get max/min value of each LC types. Broadly, LC image include forest, grassland, waterbody etc. But output max value of zonal statistic function give me 0.972 on the waterbody, 0.9908 on grassland, 0.9976 on forest. I think it is not reliable. What is the reason on this? Actually NDVI image has been masked on waterbody. Original value of the NDVI image range from -0.03 to 0.9976. Do you guys have another suggestion for this?

Thanks a lot

share|improve this question
1  
Based on many similar questions found here, I would suggest reviewing your analysis settings: if any of the input grids need to be resampled or reprojected for this analysis, that process will change the grid statistics, especially their extreme values. At a minimum, though, in order for your question to be answerable, you need to specify this information concerning your data and your analysis settings. –  whuber Jan 14 '13 at 14:39
    
Thank you for your response. But all my images are the same size and projection. –  Bandrush Barda Jan 14 '13 at 14:58
    
Do they also have identical origins and cellsizes? Regardless, your question still needs more information to be answerable. –  whuber Jan 14 '13 at 14:59
    
original cell size of LC image was little bit different with NDVI images. Then I resample it to be same as NDVI. The NDVI image has 6 bands (not just one band). –  Bandrush Barda Jan 14 '13 at 15:21
1  
That's enough to fully explain the discrepancies. The interaction of resampling and zonal statistics has been discussed in other threads here--perhaps a Google search will find them. –  whuber Jan 14 '13 at 15:26
show 1 more comment

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Browse other questions tagged or ask your own question.