This is a pretty basic question, but when calculating distance from driver variable layers, what value to you give the actual feature you are calculating distance from?
For example, if I calculate distance from agriculture, most of the cells will have a meter distance, but what should the cells within the agriculture area be valued as?
I've been using 0, but it seems to throw my model. It basically causes a lot of change to occur in those areas and that shouldn't be happening.
Hi Liz,
Distance will calculate on any zero cell from any non-zero cell. So your Ag only image does not need to be Boolean but must have zero value cells in non-Ag cells.
Also check the Metadata to make sure the units and coordinates are correct.
James
Thank you for your response! I have been using Euclidean Distance in ArcGIS and importing my .tif files into LCM.
It sounds like it doesn't matter what the value of the actual driver variable class is? Going back to my example, if I have distance from ag, the actual ag cells can be no data? Or 0? Non-Ag cells will be a distance-from value in meters, but I'm still unclear about whether the value of the Ag area matters.
That is correct, the cells you are calculating distance from can be any non-zero value. So you result should be correct if all your non-zero cells are your ag class.
Ok right now I don't have any 0 values. I'm not trying to calculate distance but make sure I am using the correct format for my driver variable in LCM.
I am inputting a driver variable into LCM that has distance in meters from a driver variable (ag). Is this not how I should be inserting the data into the model?
not sure I understand. If one of your variables is distance from ag, then when you created the distance map, you must have calculated distance from ag, so areas of ag should be zero. If not, then I would go back to that step and reanalyze. If it is, then you should be correct.
James
I am calculating distance in ArcGIS using the Euclidean distance tool and importing into LCM. This renders the ag class as nodata, but I can reclassify if needed. I'm just not sure if ag can remain nodata or how that will affect the model.
Best make it zero, and you should be fine.
Ok thanks!
Hi Liz. I see that you are also using Eucliden distance from ArcGIS to import to LCM. I am doing the same, however I get an error saying that Driver Variable 'roads' dimension does not match with the earlier and later images. Tried like a hundred times to fix things like spatial reference, extent, etc. but always this thing pops up. It would be really nice to know how you did it.
Hi Roshan - Yes I used Euclidean distance in ArcGIS and clipped to my AOI/converted to the same projection in ArcMap. I then brought the file into LCM using the GDAL conversion utility and had to use the 'project' tool to harmonize my files. Even through I used the same AOI to clip, I still had different pixel parameters.
This thread was very helpful to me when trying to figure this out (the answer is towards the bottom): https://forums.clarklabs.org/hc/en-us/community/posts/202927587-how-to-solve-dimension-problem-to-run-the-proper-LCM