Raster based monte carlo

[ToddMcDonnell]

Hello,

I am adding four different grids: RasterA + RasterB + RasterC + RasterD; and am interested in expressing uncertainty in the results based on uncertainties in each of the input grids.

All grids are in the same units and all values are positive in the range of zero to about 200.

The values in each grid have different ranges of uncertainty, such as:

+/- 6 for RasterA

+/- 10 for RasterB

+/- 12 for RasterC

+/- 5 for RasterD

I would like to run a monte carlo style analysis whereby the grids are added to together multiple times (e.g. n = 1000) based on values for each grid cell that are randomly selected from within the uncertainty range around each input grid cell value.

Does anyone have suggestions for how to carry this out or know of any available tools/code?

Thanks in advance.

[DanPatterson_Retired]

yes... I would carry out the simulation in numpy/python etc and bring the results into arcmap/Pro when needed

The first page here https://www.google.ca/search?q=numpy+monte+carlo&ie=&oe=

provides many links/videos on how to carry this out

plus searchs using Scipy include many references as well and scipy can be used within arcmap

https://conference.scipy.org/scipy2014/schedule/presentation/1699/

[ToddMcDonnell]

Thanks for the suggestion to develop with numpy/python/scipy.

I am up for developing a custom tool, but if anyone else may know of an available generic tool, I would be glad to know.

[DanPatterson_Retired]

well if you are adept at the coding domain or know someone that is, GitHub is full of such things... perhaps one of the pre-made meals might work

https://www.google.ca/search?q=github+monte+carlo+python&ie=&oe=

[ToddMcDonnell]

Thanks

[ToddMcDonnell]

I would also be interested in additional solutions/suggestions that may be available for carrying out such an analysis.