nav[aria-label="Primary Navigation"] { padding: 0; & ul { list-style: none; width: 100%; display: flex; flex-direction: row; justify-content: start; align-items: start; gap: 30px; padding: 0; & li { margin: 0; } & ul li { list-style: none; } } }

ArcGis Pro Zonal statistics as table

Hope-Hauptman

I have a soil classification raster that I clipped to a 500m radius circle around central contamination points. I am trying to determine the majority soil type (hydrologic group) in the clipped rasters using the Zonal statistics as table tool but I must be doing something wrong?

I have done this successfully with another raster but there was only one value in the attribute table to calculate the majority within the clip, here there are a more than one column in the attribute table.

For example in the picture - visually most of the buffer zones seem to have a majority blue by area (soil type but I would like ESRI to calculate the majority area for each circle for me...

Capture1.PNG

CA_MURASTER_SJV_Clip.lyrx

BatchIntersect.png

Append.zip

MarkusSchenardi_0-1715955130725.png

Bud_1-1715955005695.png

Find more posts tagged with

Analysis

Desktop

Accepted answers

curtvprice

> I did do this for another raster but there was only one value and it was categorical - landuse type.

If you want to do it that way, the solution is to use the Reclassify tool to convert your soil types into integer codes (1,2,3,4..) (like your land use codes were), do your zonal statistics as table to find majority value, and then map the results (1,2,3,4..) back to soil codes using an expression or lookup table.

I know this seems unwieldy but is because ArcGIS rasters do not store text, they store integer or float numbers, so to do calculations on categorical data the values must be integer, not text codes like "A/B".

All comments

curtvprice

Remember, you can only do statistics on number codes not text codes, that is the root of your issue.

When I look at your raster I think the easiest approach is to use the Lookup tool with your raster as input, and specify the text field HYDROLGRP_DCD. This output raster from the Lookup tool will have a unique Value field (an integer) and your text field in it. Run your Zonal Stats As Table on this using the Value field. Then you'll have the majority soil code (as the Value integer). To add the soil text code to the zonal stats table, use the Join Field tool with your zonal output as the input table, input field Value, the raster table (with Value and HYDROLGRP_DCD) as the join table, join field MAJORITY. This will add your soil codes into your zonal stats results.

JayantaPoddar

I don't have a similar data to test. Could you add just the buffer layer and the Clipped Raster layer on map, and create a map package? Share the map package through a Google Drive, Dropbox or FTP link.

Let me have a look at the data.

Hope-Hauptman

Great question. Last month I also used Zonal Statistics as Table tool to calculate majority land use but the VALUE field was the land use code (see screenshot last month). Now with my current problem the VALUE field is not what I want to calculate my statistics on. I want hydrologic soil class. Do I need to title it VALUE instead? Another suggestion was to use the lookup tool but that did not work... Not sure what I am doing wrong. Sorry I know this is a repetitive question and I thought I had this problem solved last month...

last_month.PNG

current.PNG

Getting_Data_Science_with_R_and_ArcGIS.pdf

1.Instalación.pdf

curtvprice

Jayanta, the difference is now the input zone values are text, not integer (see my previous post). The land use codes were integers (Value field), the soil type values are in a text field in the raster table.

JayantaPoddar

Something is missing. How is your current requirement different from ArcGIS Pro determine dominant land use type around each point in a given area (500m buffer) that you posted last month?

curtvprice

One more thing this hydro groups are categories and if your interest is to look at well risk, this may be best handled for this analysis by converting the cell values to a vulnerability score (say 1-10) and doing an average instead of picking the majority. Just a thought.

curtvprice

Rasters are stored and manipulated as numbers not text. So to get the value in the table you probably need to do a join (Join Field is probably the easiest way to do this) to copy the hydro group value over to the table from the input raster table (Value being the join field).

Another thing to look out for is that the majority is not necessarily a plurality, that is if a buffer has enough hydro groups, you may want to use the Tabulate Area tool instead so you can see those special cases and deal with them. This will involve a little more database work on the output, but I think it is probably worth it.

Hope-Hauptman

Yes check on all those boxes. I tried it again using the full raster as the input raster but instead of the majority soil hydrologic group as the majority group I get an area number for the majority? See pic.

Capture3.PNG

R-ArcGIS-DevSummit-2019.pdf

JayantaPoddar

A few check-points. Please verify

1. Use the buffer polygon layer as Input Feature Zone Data. For the Zone Field, use a field of unique values like "ObjectID". Ensure no "Buffer" polygon Feature part is selected in the map.

*EDIT: Ensure there are no multipart feature in the Buffer layer. While creating the buffer from the points, the Dissolve Type parameter should be "No Dissolve".

2. For input value Raster, you may use the clipped raster (or even better, the full Raster).

3. For Statistics Type, check with "MAJORITY"

curtvprice

I did build a tool to Weed Points (NACT) so you can do them in groups. But, I looked in the help and at ArcGIS Pro version 2.8 (but not ArcMap 10.8, unfortunately), Zonal Statistics As Table handles overlapping polygons. How wonderful!

I re-read your question. If the value is in another field in the raster table besides Value, you can create a new raster of the table value you want using the Lookup tool.

Hope-Hauptman

This is just a small picture. I have thousands of clips.

curtvprice

Your buffers overlap. To get correct results you need to run the Zonal Statistics (or Tabulate Area would be better so you can see more detail) tool for each circle one at a time; you could do this using modelbuilder or python iteration. There are some development efforts to avoid the overlap problem but since this is a small number I would just do the buffers one by one.

Moving this to the Spatial Analyst space where lots of smart people may have more suggestions.

Join the Conversation

Signed in members can post, follow updates, and more. New here? Register a free account.

Quick Links

Community Help
Find useful guides, FAQs, and documents to help you navigate and make the most of Esri Community.