Modelling for species or community interactions at timepoints

30 Views Asked by At

I have data collected of abundances, mixed taxa species for three different years where each row of the data frame corresponds to a species observed at a unique site. I collected arthropods from trees at specific time points: 0.43, 1, 2, 3 weeks since previous observation. The data contains missing categories of these time points, but I would like to see if there could be some relationship with the arthropods seen at one time point to the next. Like for example, anecdotally, I saw a tendency for spiders to eat some insects and whenever they occurred the other species either didn't or in smaller numbers than the previous timepoints of observation. Would there be a way to model how a community changes over time? Or would there be a way to look for patterns in the species-time data that might explain why some communities appear diverse and others not? I think I would need something like a random forest model and for my data to be at the site level, and maybe like a distance matrix. Any direction would be super helpful! Thank you, M

1

There are 1 best solutions below

0
proffarnsworth On

I'll point you to these two papers:

https://link.springer.com/article/10.1007/s10336-016-1404-9

Accounting for missing data when assessing availability in animal population surveys: an application to ice-associated seals in the Bering Sea

They both outline a statistical approach for filling in missing values in population data. I understand that you are suggesting to fill in missing values based on known population of spiders. This could be done with random forest or a similar regression model, but there may be stronger statistical approaches to explore before going that route. I would also suggest creating a visualization of the data you currently have to explore correlations in the data.

Without any data or code provided though I can't suggest a solution in R.