Getting rid of high values in Shapefile Polygons

Question

Hey Guys,

I´m pretty new to GIS and currently struggeling with my masters thesis, so I hope you can help me out.

The Problem: I have a few debris flow simulation shapefiles which consist of up to 200,000 1m x 1m square polygons. Each polygon has its own value of deposition height. Unfortunately there are some unrealistic high values in some polygons, due to the simulation of the debris flow (probably less than 0.1% of the whole amount of squares). That makes it difficult to show the deposition height of the flow properly in color. I want a green to red colorscheme but red colors are not really shown in the map because there are too less values that grow so high. So most of the deposition remains green or yellowish. There are too many to pick them out by hand.

Is there any chance to get rid of these unnaturally high values e.g. reducing the resolution or use some nearest neighbour method too smooth out the distribution of the values.

I hope you can help me

DanPatterson · Answer

Add a new field of the appropriate type.

Query !oldfield! > 199999      for instance where the big number is the the upper threshold of what you want to toast

Switch the Selection!  ..... only the rows that contain good data will be calculated

make the newfield active... field calculate    !oldfield!

only the selected records will get the appropriate value that you want to keep and the "bad" data will be null

Approx time under 2 minutes

NewGuy · Answer

Thanks for your answer.

It´s basically just for visualization, so it won´t affect any further evaluations.

I was already thinking about using the field calculator to filter out values which only occur in small percentages but thats really complicated an time consuming. I really want to avoid that but if theres no other option like the resample tool for rasters, i guess i´ll have to go with the calculator method 😞

DanPatterson · Answer

You can Select Layer by Attribute using the field containing the deposition height based on some threshold (you need a histogram for that.  You could then remove them to their own layer or you could calculate a new field in the existing file, setting their value to Null (aka None in the field calculator) and copy over the "good values".  Use that field for symbolizing.

Or you could convert everything to raster and there are lots of options then.

What is your end-game? is it for visualization? or for analysis?  It will have an influence on how to proceed