Models without spatial information

Summary

• Repeated the analysis here http://rpubs.com/Chopin/412588, but without spatial position information.

• Without lon and lat, P and %N met the requirement of 5 % relative influence in the BRT models.

• Overall, the models had similar, or marginally worse predictive power than the models with spatial information.

• The major difference was that fewer covariates were significant in these models.

Identifying important covariates

Redid the variable choice analysis in http://rpubs.com/Chopin/412588 but with lon and lat excluded. P and perc_N showed up as having a relative influence of 5% in at least one of the BRT models. The final variables used are below.

 fin_occ_df <- dat_occ %>% 
  select(1:10, 
         ph_et_al,
         salt,
         carbon,
         Ca,
         Q_cover,
         elevation,
         percent_over1,
         percent_over2,
         P,
         N_perc,
         aspect,
         texture,
         corr_dN,
         drainage)

Fitting the JSDM

The model fit looked good with scaled covariates.

## NULL

Predicting with JSDMs

The above model was used to predict onto the same data. The AUC values were all close to one, suggesting a very good predictive power.

Species	AUC
R_burtoniae	1.000
R_comptonii	1.000
D_diversifolium	0.999
A_delaetii	1.000
A_fissum	1.000
A_framesii	0.999
C_spissum	1.000
C_staminodiosum	0.988
Dicrocaulon_sp	1.000
Oophytum_sp	0.997

Let’s use sites 2 and 3 to predict site 1. Some species are predicted really well, others very badly.

Species	AUC
R_burtoniae	0.911
R_comptonii	0.707
D_diversifolium	0.357
A_delaetii	0.910
A_fissum	0.415
A_framesii	0.665
C_spissum	0.357
C_staminodiosum	0.729
Dicrocaulon_sp	0.359
Oophytum_sp	0.917

Now sites 1 and 2 on 3. Predictions are much worse than they were. A_framesii is not included as it didn’t occur in site 3. It’s a bit strange that the auc values are exactly the same as there were for the last analysis, but I can’t find and error in the code.

Species	AUC
R_burtoniae	0.851
R_comptonii	0.518
D_diversifolium	0.426
A_delaetii	0.575
A_fissum	0.417
C_spissum	0.492
C_staminodiosum	0.725
Dicrocaulon_sp	0.486
Oophytum_sp	0.660

Now sites 1 and 3 on 2. Predictions are much worse than they were. R_comptonii and C_stamin are not included as it didn’t occur in site 2.

Species	AUC
R_burtoniae	0.943
D_diversifolium	0.323
A_delaetii	0.682
A_fissum	0.078
A_framesii	0.526
C_spissum	0.402
Dicrocaulon_sp	0.635
Oophytum_sp	0.650

Let’s try by randomly selecting 100 plots and predicting onto the remaining 50. It predicts a lot better than site by site.

Species	AUC
R_burtoniae	0.870
R_comptonii	0.960
D_diversifolium	0.778
A_delaetii	0.836
A_fissum	0.435
A_framesii	0.833
C_spissum	0.619
C_staminodiosum	0.780
Dicrocaulon_sp	0.841
Oophytum_sp	0.951

Overall then, there is some indication that distribution is deterministic.

Abiotic or biotic?

Boral can partition variance into that explained by environment and latent variables. However, this seems quite sensitive to the data used. The full model is shown on the left, and the randomly chosen one on the right.

Correlations

Using the full model. Correlation due to environment on the left, correlation due to latent variables on the right.

---

Using the random model.

Important covariates

Full model