Spatiotemporal analysis of binary (presence/absence) data

[AnneProvencher_St-Pierre]

Hi ! 

I have a dataset which is in the form of a grid, where I identified the presence or absence of a species at each grid intersect (about 10 000 in total), and this was repeated for 6 time points. 

I recently learned about space-time cubes and emerging hot spot analyses and thought it would be a great way to explore my data, but it seems that this cannot be applied because I do not have enough temporal layers (i.e. the Space Time Cube tool asks for a minimum of 10). Is that the case ? Is there anything I could do to go around that issue, or similar analyses that I could apply ?

Also, in addition to the presence/absence data I also have values for environmental parameters for each grid intersect. These parameters do not change over time. I'd like to apply an analysis to look at which parameters might explain why the species is always present in an area vs always absent in another, etc. Any suggestions ?

Thanks a lot !

Anne

[DanPatterson_Retired]

begin with a 'sum' using cell statistics on your raster stack

Cell Statistics—Help | ArcGIS Desktop 

A quick way of looking at the time thing beyond a sum of 0, 1, 2, 3, 4, 5, 6 is to class you first raster as 0,1, then your second, multiply by 2, the third, multiply by 4, … 4th, by 8, 5th, 16, 6 by 32.

So you will have a progression of 0, 1, 2, 4, 8, 16, 32 at a location at each point in time. so if a site has a value of 63, it was a 1 for all time periods... a value of 48 shows a 1 for the last 2 time periods etc … all using cell statistics, although the raster conversion and summations are quite fast in numpy

/blogs/dan_patterson/2018/02/06/cell-statistics-made-easy-raster-data-over-time