Number of species per group

The distributions of our grouped richness data are displayed on the below histograms. The red line is overall sample mean and the blue line is the median:

We can see that the number of insects counted per group is slightly skewed (positively), but not by much in most cases. Notably, due to having a very small number of wood-modifying species (bark borers), the distribution for this group is heavily zero-inflated.

Note ants are not included due to the inconsistency of species identification mentioned in Document 01*

Next we’ll look at how number of species (standardized by survey time) correlates to various traits of import:

There is no discernable relationship between PGs and group richness.

A similar pattern occurs with CTs.

Not surprisingly, there is a negative relationship between species count density. However, it is important to remember that the time spent searching for insects on a given tree was a function of the tree’s height.

Therefore, volume and count data are confounded. Below are plots of the total number of species, this time NOT accounting for survey time, found per unit volume:

Here is a heatmap of the correlations for some of the above relationships:

We can see that most correlations are pretty weak.

Number of individuals per group

We can see that the individual data are very positively skewed due to zero-inflation. Therefore, we will look at the relationships excluding zeros for now. i.e. \(Y = Density | Present\).

Removing all zeros only slightly improves the situation.

There is not an obvious relationship between group density and PGs

Nor is are there very strong relationships with CTs

There doesn’t even seem to be a relationship with volume.

Here is our correlation heatmap, this time using density data:

Only weak correlations exist between traits and group density (including zeros)

Next we’ll check for interesting among-group correlations:

There don’t appear to be any interesting group-wide correlations.