I am doing dissertation about [predicting future urban expansion for Shannon River Basin In Ireland using Cellular Automata model and arcGIS] so at this stage I would like to calculation the conversion probability factor (Land use transitional probabilities), as it is advised that the value of this probability can be generated fixed by overlaying two land use maps, so it can see the value of conversion rate of a certain land use to other land use. so that I enter it into the model.

I am kinda stuck on which arcGIS tool of OVERLAYING should I use (union, erase,identify etc), knowing that I have two landuse/landcover maps from EPA website ready with almost 32 different landuse classes for the years 2000 and 2006 and how can I actually calculate the conversion rate ?

closed as unclear what you're asking by PolyGeo Aug 1 '17 at 5:15

Please clarify your specific problem or add additional details to highlight exactly what you need. As it's currently written, it’s hard to tell exactly what you're asking. See the How to Ask page for help clarifying this question. If this question can be reworded to fit the rules in the help center, please edit the question.

  • Does CA stand for Cellular Automata? – fatih_dur Jun 25 '15 at 23:08
  • Yes sorry ill edit that – Sarah Jun 25 '15 at 23:13
  • OK, I think there are other considerations that you need to clarify before starting to calculate desired outcome of this exercise however we can discuss the basic of this query. The tools that you are intending to use are vector overlay analysis tools (i.e., your land use data is most probably supplied in vector format) and at the first stage it is OK to use however in the later stages you will be using raster data for CA part. This transition from vector to raster may make things a bit complicated. Another approach converting all land use data to raster to start with and carry on analyses. – fatih_dur Jun 25 '15 at 23:37
  • Basically the latter approach is to make sure you will be using the same unit of analysis throughout. If it is going to be an auto-regressive model (disregarding exogenous and endogenous factors), you may employ Spatial Analysis toolbox, specifically raster calculator and focal statistics together. By this you can produce n-->n+1 probabilities for each cell and then you can find a joint probability for 5 periods. It's been a while that i am not dealing with spatial analysis and statistics matters, therefore these are my initial thoughts to give you an idea. – fatih_dur Jun 25 '15 at 23:42
  • 1
    Lastly spatial analysis's focal statistics tool uses the neighborhood concept to produce stats and you need to find right definition for that (similar to CA). Also people use other means for the similar problems (i.e., R or other Python based modules) considering poor performance of ArcGIS in spatial analysis or pure performance /control considerations. – fatih_dur Jun 25 '15 at 23:45