0

CA Markov module: How long does processing take?

Charshade17

April 04, 2018 18:59

None

Follow

Hello! I am currently using the CA_Markov module in TerrSet 18.3 to create a land cover of the city of Philadelphia for the year 2044. As you may imagine, my rater for the city is quite large, and it is broken into 8 land use categories. I input a LULC raster from 2015 into the module, so it is running 29 iterations (1 iteration for each year from 2015- 2044). The module has been running for ~24 hours now and it is still on pass 7 out of 1624. Any suggestions why this may be? All of the data is stored on a 2 TB external hard drive. Also, any idea form CA_mola means (see picture)?

Thanks!

 

• Facebook
• Twitter
• LinkedIn

1 comment

• 0
  
  Charshade17 April 05, 2018 19:26

  Hello,
  
  CA_Markov is a fairly intensive module since it is based on cellular automata for making the predicted land cover map. The size of the basis land cover image will play a large role on how long it takes to run. How many rows and columns does your LULC raster from 2015 have? Also, the other input parameters for this module will influence the time it takes to run. What did you input for number of CA iterations and the filter size?
  
  You might want to progressively double check your task manager to make sure that the module is still running and has not crashed. The specs of your computer will also influence how long it will take to run. One major factor that will make it run slow if the fact that the data is on an external harddrive as appose to on the computer itself, especially if it is a large dataset like you mentioned. If you can give up some spatial resolution, I would suggest using the contract module to make the imagery a bit coarser, as this could greatly speed up run time.
  
  Lastly CA_mola refers to the Multi-Objective Land Allocation module that is called behind the scenes.
  
  Hopefully this was helpful.
  
  Best regards,
  Tom